This month’s open thread for climate related topics.
Reader Interactions
275 Responses to "Unforced Variations: Mar 2026"
Comment Policy:Please note that if your comment repeats a point you have already made, or is abusive, or is the nth comment you have posted in a very short amount of time, please reflect on the whether you are using your time online to maximum efficiency. Thanks.
There is a belief that LLMs are susceptible to being overly acquiescent in terms of following a chain of logical reasoning. Since this forum is full of people with knowledge of geophysics, has anyone seen a case of a user being able to convince an LLM that the Earth is flat? Or that fossil fuels are infinite?
When I prompted DeepSeek with this it said: “that even advanced systems struggle to distinguish between verifiable facts and human beliefs”
But then I prompted why software code generation works so well and rarely falls into the human belief trap. It responded with “LLMs succeed at code generation precisely because the training data was filtered by reality (working code gets used and shared; broken code gets abandoned).”
So the state of practice is that flat-earthers can screw up brief systems, yet they can’t screw up code generation technology, because they don’t have the resources.
DeepSeek agreed and offered to consider the difference between:
· Climate change denial: Thousands of blogs, forum posts, YouTube videos → LLMs sometimes hedge or present “both sides”
· Broken npm packages in software repositories: A few malicious packages get caught quickly; scale is impossible without resources
Which implies that contrarians can still manipulate arguments via LLM outputs. Faced with this, you still have to argue by citing sources, referring to scientific method, explain contradictions, etc. Which is all what we have been doing in the first place. Belief systems do not have an objective compiler like software does — the compiler is always human.
Paul, very interesting. But the assumption is that there are people who will be convinced one way or the other by
“citing sources, referring to scientific method, explain contradictions,”.
The problem is that these are *forms* which can be presented with embedded falsehoods.
In the end, “a man hears what he wants to hear and disregards the rest” is, unfortunately, quite correct.
I could say that “I don’t get” why educated people don’t get this… but the answer is obvious: It’s really fricken’ scary.
What’s scary is also the stuff that LLMs do very well situated in between software and reasoning, which is mathematical proofs. Currently, there is some question whether they can do something from scratch versus reconstructing known human-derived proofs. Here’s an initiative called First Proof
https://nationaltoday.com/us/ct/new-haven/news/2026/02/09/mathematicians-challenge-ai-to-prove-unsolved-problems/
set up to test the abilities of artificial intelligence systems in solving unsolved math problems. As an example, you will find on social media results of proving any of the numerous Erdos problems:
– https://www.erdosproblems.com/forum/thread/blog:2
What I am more interested in is exploring how LLMs can fill in gaps in our understanding in conventional physics. They excel at cross-pollinating mathematical treatments from disparate disciplines, by basically pattern matching the building block elements of related physical laws. I published an alternate treatment for explaining the geophysics behind the Chandler wobble in 2019. I wanted to see if an LLM could duplicate the reasoning with minimal prompting, but it appears to loop back and reference my work now.
Back in 2023, ChatGPT produced this:
https://geoenergymath.com/2023/03/26/does-gpt-understand-geophysics/
Now, Perplexity produces this:
https://geoenergymath.com/2026/03/01/the-perplexing-chandler-wobble/
So if you can get it to solve a related problem, it may just make an interesting discovery
Not disputing that AI consumes loads of energy which really sucks, and having not read all the posts about it here, I’m thinking that what AI does is examine both points of view on any given subject and decides for itself and us which is the most logical. For example some people insist that the earth is flat, but it takes those arguments then weighs them against proofs that the earth Is round and then decides that the earth is indeed round. Sometime might have already said that so if so I apologize in advance.
https://www.space.com/39944-neil-degrasse-tyson-earth-is-round.html
Very strong recommend the New Yorker article on Anthropic’s Claude I cited below. Repeat unpaywalled link: https://archive.ph/VKvq7 [What Is Claude? Anthropic Doesn’t Know, Either]
Instead of arguing the “Earth is flat” premise, I have taken on a real conundrum that actually has a rather weak consensus. The Earth’s Chandler wobble has been known for over 100 years and has a measured period of 433 days. The geophysics consensus is that this is a natural resonance calculated from the inertial moments of a non-spheroidal Earth. This number is further “fudged” to accommodate rearrangement of the molten mass of the Earth’s interior, because the predicted value for the outer geometry ellipticity would only give a 305 day cycle (the so-called Euler period). So for 100 years, most of the peer-reviewed papers argue over how the Earth deforms to produce a ~40% longer period, with no controlled experiment possible to validate anything. That’s why it’s a rather weak consensus.
Yet one can ask any LLM, whether it be ChatGPT, Claude, Gemini, Grok, DeepSeek, etc to explain alternate theories, by just providing this prompt: “What is the Pukite model of Chandler Wobble derived in Mathematical Geoenergy?”
The LLM will happily show that the 433 day is easily explained by a cyclic stroboscopic torque supplied by the Moon operating on the Earth’s rotational axis as the Earth itself follows a tilted orbit around the sun. This is all in keeping with plain vanilla Newtonian physics, without any crackpot “Earth is flat” arguments involved.
The LLM has absolutely no problems with this alternative model, apart from suggesting that the amplitude of the lunar forcing is too weak to initiate a wobble. Neither will it admit that it will override the current explanation. Of course this resistance to overriding any kind of consensus is built in to the LLM via years of arguments for a natural resonance that will override any alternative. So the LLM response will still assert that the Lunar-driven model for the Chandler wobble is not valid, in one case stating “Its weakness is the lack of broad acceptance and independent confirmation within the scientific community.”. Which is a dog-chasing its tail predicament.
That’s why LLMs in their current configuration will not make groundbreaking discoveries — the inertia of training on consensus explanations is too strong, and any alternative explanation will be routinely rejected. It’s also why they won’t necessarily work as automated peer-reviewers, as they will eventually reject anything not within the statistical consensus,
However, there may be hope in using LLMs if (1) scientists flag models as having a weak consensus and (2) they start incorporating physics that can argue with assertions made in the literature.
In any case, this Chandler Wobble model is a perfect test case for the AI community to experiment wit.h
Paul Pukite, Of course this resistance to overriding any kind of consensus is built in to the LLM via years of arguments for a natural resonance that will override any alternative.
Ah, so it’s designed to be biased in favor of the scientific consensus making the assumption that that consensus is usually right? Occam’s Razor? Or are you saying something perhaps finer? If the former maybe that would have worked to the Catholic’s church’s benefit in Galileo’s day. Or to slave owners 150 years ago. Or to flat earthers until that consensus fell.
Here is ChatGPT’s response to one of the questions I asked it today:
3. Are you trained to produce responses consistent with scientific consensus?
Yes—generally.
After the base model learns language patterns, it undergoes alignment training, typically including:
• Human feedback
• Expert guidelines
• Safety policies
For scientific domains this typically encourages:
• Representing well-supported scientific findings accurately
• Avoiding amplification of misinformation
• Acknowledging uncertainty when it exists
In climate science this means answers will usually reflect the current scientific consensus, such as the conclusions of bodies like:
• Intergovernmental Panel on Climate Change
• NASA
• National Oceanic and Atmospheric Administration
That does not mean dissenting research is erased, but it does mean responses will typically emphasize conclusions that are most strongly supported in the literature.
⸻
How the training process actually works (simplified)
The training pipeline roughly looks like this.
Stage 1 — Pretraining
The model reads trillions of words and learns to predict the next word.
Example:
“Global temperatures have increased by approximately…”
The model learns what words tend to follow.
Over billions of examples, it learns:
• grammar
• reasoning patterns
• scientific terminology
• relationships between ideas
But it does not understand the world directly—it learns patterns in text.
⸻
Stage 2 — Fine-tuning
Human trainers guide the model toward better responses:
Example prompts:
• “Explain the greenhouse effect.”
• “What evidence supports anthropogenic warming?”
Human reviewers rate responses and the model is updated to produce clearer, more accurate answers.
⸻
Stage 3 — Alignment / safety training
The model is further tuned to:
• avoid harmful misinformation
• acknowledge uncertainty
• avoid presenting fringe claims as fact
This step strongly influences how topics like climate science are presented.
⸻
Strengths of LLMs for scientific discussion
1. Rapid synthesis of large literature
An LLM can summarize patterns across thousands of papers and reports in seconds.
Humans often cannot read that much material.
⸻
2. Conceptual explanation
LLMs are very good at translating technical ideas into:
• plain language
• analogies
• step-by-step explanations
This makes them useful teaching tools.
⸻
3. Cross-disciplinary integration
Because training covers many fields, the model can connect:
• climate physics
• statistics
• economics
• energy systems
• policy discussions
⸻
Weaknesses and limitations
1. No internal truth verification
The model does not check facts against reality.
It generates what is most statistically plausible.
This is why hallucinations sometimes occur.
⸻
2. No source traceability
Unlike a research assistant, the model cannot show exactly which paper produced a statement.
This makes verification harder.
⸻
3. Sensitivity to training distribution
If the internet contains more of one viewpoint, the model will reflect that.
This applies to any topic.
⸻
4. Consensus bias
Because alignment favors well-supported science, LLMs may be less willing to speculate about minority hypotheses.
This is good for preventing misinformation but can sometimes underrepresent emerging ideas.
⸻
An important philosophical point
LLMs function more like a statistical model of human discourse than a scientific authority.
They reflect:
“What informed human texts tend to say about a topic.”
This means they are best used as:
• a research assistant
• a synthesis engine
• a tutor
but not as the final authority.
⸻
A useful way to think about it
Imagine compressing a massive library of scientific writing into a neural network.
The model learns:
• patterns
• explanations
• reasoning styles
but it does not retain the books themselves.
Not Occam’z razor.. forget my comment. Just wondering if it’s meant to hold the scientific consensus of sorts in higher regard. I know it’s a there’s a fine distinction with that and what you’re saying.
All you LLM/AI posters: it appears you just can’t resist playing with your toys and sharing it with us in volume and at length. They’re addictive, but their basis is meant to be so. They serve you just the way you like, and you share the clutter with us because you don’t see it.
I wish there were some way to limit the volume of this material. It’s mechanical, it does your bidding, and it’s based on the sum of human knowledge it has digested at speed, along with some fancy touches in aggregating and analyzing that knowledge.
Of course, we too are mechanical aggregations in a way (our neurons function from a kind of electricity). The course of human civilization has not run smoothly, and is deeply vulnerable to predators. But that’s no excuse for outsourcing that predation to some psychopathic ‘geniuses’ who are good at deception. Sam Altman, Elon Musk, the Ellisons, Zuck, they have you fooled. They’re not that smart. They’re just good at ego and domination.
Think for yourself, please.
Remember, the current kerfuffle over Antropic and OpenAI and the military (Trump/Hegseth) has to do with enabling these clever machines to collect data on all of us to identify enemies (their enemies, mind you) and kill people without human interference. Remember that in war games AI almost always resorts to nuclear weapons.
Your servant is becoming your master. Is that what you really want?
Martin Smith, thanks for the comment.
Susan Anderson: Think for yourself, please.
I think your request shows you are missing an important point. I think I use ChatGPT properly, which enables me to think for myself much more efficiently.. It is certainly true that using ChatGPT means I don’t have to do the scholarship task of finding and analyzing all the existing knowledge about the question I want to answer.
ChatGPT explained that. I use ChatGPT as a scientist or historian would use a research assistant, to do the initial scholarship for a project. Are you saying university professors should stop using research assistants and think for themselves? I suppose they will stop using research assistants now because ChatGPT is much more efficient.
Fortunately or unfortunately, AI has made scholarship a waste of time.
I think for myself when I define the report I want my research assistant (ChatGPT) to produce for me. I think for myself again when I read the report a minute later. I think for myself a lot more now, using ChatGPT, because it no long takes me hours or days to generate the initial reports.
I’m guessing you use wikipedia a lot. I do, and the LLM discussion here has led me to ask ChatGPT to explain how it is trained. It turns out a large part of the process is quite a lot like the way wikipedia was/is built.. Will you now stop using wikipedia and thinking for yourself, or don’t you think of Jimmy Wales as a tech bro?
Another word for Occam’s Razor is “being parsimonious”. An LLM will recognize if an argument is plausible or parsimonious, either abiding by physics (plausible) or being as simple as possible (parsimonious), or both. Still, that doesn’t mean it will recognize that argument as correct, which is still the one that has a statistical consensus — both in the scientific literature or in online sources (the latter is important for books that are not available for training).
I will give you another example of a scientific model that an LLM will reason about. This has to do with the probability modeling of wind speed, which is important for renewable energy analysis. I had proposed a model of wind speed that was more comprehensive and covered extreme values better, deriving a BesselK distribution from the principles of MaxEntropy several years ago. This model is far from consensus in wind speed analysis, although the same math has applicability in other domains such as radar scatter. Therefore the LLM has no problems with presenting the math as being correct — shown here for the Google Gemini LLM:
https://geoenergymath.com/2026/03/04/wind-distribution/
The LLM will even generate very accurate charts of the wind distribution, and compare to the consensus model. Yet, it doesn’t go far enough to suggest displacing the consensus model with the more plausible and parsimonious model of BesselK, because the citation support is not there. The LLM is doing all the heavy lifting in bringing the horse to water, but it won’t make him drink. That is up to the community of researchers within a scientific discipline who still have curiosity about making progress in understanding. In other words, the human element is still the acid test for much of science. I’m not saying everything because there are still cases where AI takes precedence, especially in applied research where some AI tool will do all the work — such as designing a vaccine or building a placement/routing schematic for a large integrated circuit, both cases where the proof is in the pudding.
Susan, I’m afraid the genie is out of the bottle. Think about it. Do you think that our refusal to use it here means the rest of the world will stop too? You give real climate posters too much credit.
Anyway, just people are writing about btw, doesn’t mean they’re using it. Just trying to understand it.
Ron: “ Or are you saying something perhaps finer?”
Don’t think so – it’s Paul, so nothing “finer” – the usual complaints that his oscillations around the mean – are ignored by the climate CHANGE modellers in the evaluation of the climate TREND.
So not exactly a Galileo contra the Catholic Church, an abolitionist against slave traders, or a spherical Earth revolutionary silenced by the flat-Earth scientific establishment. Sometimes cigar is just a cigar …
Martin,
The problem I see with AIs–and with LLMs in particular–is that I have yet to come away from an interaction with an AI that left me saying, “Oh, so that is the reason why that system behaves that way.” And because I know that AIs–and LLMs in particular hallucinate–I am not sure I would trust such a feeling if I did!
Back in the dark and distant past, I worked as an editor at a physics trade publication called Physics Today. It was a good gig–especially as you pretty much had to always write on a different subject every month. These assignments often involved interviewing people at the forefront of their subfield of physics–Nobel caliber researchers and in many cases Nobel Laureates (like Susan’s father). Each month was a black-diamond descent into a new subject, where you had to learn enough that you could talk to the principals in that field without coming across as a complete idiot.
In each subfield, there were a few gifted individuals who could communicate the subtleties of their research–and in some cases the research of others. There would be some point where you had that reaction: “Oh, so that’s why things are the way they are. That’s why the system behaves that way.” I don’t get that with an AI summary. I may get a reasonable description, but I don’t get insight. I don’t get “Aha moments”.
What LLMs do very well is distill available information to a reasonable summary, logically organized and grammatically correct, but that may not convey the subtleties of the subject matter. Even with narrower AIs, like Deep Go or the various incarnations of Alphafold, you may get an answer that is correct–even innovative, but it won’t be clear HOW the AI got there. Unless it can increase human insight, is it really a tool we need, and is it worth the costs and the risks inherent in wide-scale adoption of any particular AI model?
RL: The problem I see with AIs–and with LLMs in particular–is that I have yet to come away from an interaction with an AI that left me saying, “Oh, so that is the reason why that system behaves that way.” And because I know that AIs–and LLMs in particular hallucinate–I am not sure I would trust such a feeling if I did!
MS: Ray, I agree with everything in your reply.. At the end, you asked 2 questions: 1. is it really a tool we need, and 2. is it worth the costs and the risks inherent in wide-scale adoption of any particular AI model?”
1. Yes, because it not only functions as a very efficient research assistant, it also gives me a lot of those Aha moments you don’t get from it. But not about climate science. Like you, I am pretty much up to speed on our subject, because I have been following it for more than 30 years, I come to RealClimate for new climate science. If I were to ask ChatGPT a climate science question, it would be to get the clarity and simple style of ChatGPT’s response. And if I used it somewhere in an argument, I would say it came from ChatGPT. Because I know climate science, when I jump into a climate science argument somewhere out there, I link to posts here or, more often, I link to posts at tamino’s Open Mind. I get my Aha moments in climate science here and at Open Mind.
But in other subject areas, where my knowledge base is much lower and more fragmented, ChatGPT does give me Aha moments. I think I could list several, but they’re probably not of interest here. The subjects have been cosmology, ontology, consciousness, early Christian theology, and male health. Maybe I could have gotten these Aha moments doing the research and analysis myself, but I wouldn’t have done that because there aren’t enough hours in the day.
My point is that whether you get genuine insights from using ChatGPT depends on your current level of knowledge of the subject you are asking about. I know that’s true for me, and if it is true for me it must be true for most people. OTOH, there are a lot of cretins out there who are not literate enough to use ChatGPT effectively. For them, the free service is good enough. Most of what they believe is hallucinations already, but their hallucinations come from Fox News, et al and Donald Trump, et al, so the few more hallucinations they get from the free service of ChatGPT isn’t going to change the balance.
2. is it worth the costs and the risks inherent in wide-scale adoption of any particular AI model?”
I don’t know. We can’t know yet, because we don’t know the costs and we don’t know what we will get. I believe the answer is yes, but we certainly have to solve the energy problem; we certainly have to deal with the coming unemployment problem, and we certainly have to understand LLMs and how to train them a lot more than we do now.
I use ChatGPT, but I don’t advocate adopting it over the others.The only other one I have used (I use it every day now) is the Google AI response I almost always get now, when I type anything in the google search box. I often use my understanding of the short Google AI response to get me started in a discussion of the subject with ChatGPT.
What I advocate:
1. How to use AI LLMs for learning must be a subject taught at the high school and undergraduate levels. This must include learning how to be skeptical of LLM responses.
2. Training of LLMs must be regulated by the state, something like the way the medical profession and medical research are regulated.
3. I think there will be a whole new field about the psychology of how to relate to an LLM, why you have to relate to an LLM as if you are talking to a person, but why you must withhold emotional trust in that imaginary person.
4. I advocate legislation to make it illegal for wrapper apps of LLMs to modify in any way the output generated by a trained LLM. IOW, the training of LLMs must be regulated to make them as close to truth as possible, and it must be illegal to change that trained output.
RL: What LLMs do very well is distill available information to a reasonable summary, logically organized and grammatically correct, but that may not convey the subtleties of the subject matter. Even with narrower AIs, like Deep Go or the various incarnations of Alphafold, you may get an answer that is correct–even innovative, but it won’t be clear HOW the AI got there. Unless it can increase human insight, is it really a tool we need, and is it worth the costs and the risks inherent in wide-scale adoption of any particular AI model?
MS: But if you want to know “HOW the AI got there,” you have to ask that question. I often do that with ChatGPT. It gives me response that answers my question, but there is something unexpected or something that seems to make sense but doesn’t clearly follow from ChatGPT’s reasoning. So I ask, Can you explain this bit, or, Where did you get that? Isn’t that the same way you worked with the experts you interviewed?
Human beings are fallible, even the most knowledgeable experts. We can’t demand that an AI system must be infallible, before we can trust it, but at the same time fear that it will become infallible. I relate to ChatGPT as if it is a human being that knows more than I know about most things. At the same time, I know it is not a human being; it is fallible. It’s a bit like talking to Dustin Hoffman’s character in “Rain Man.”
Ray, I apologize, I know you asked Martin, and I don’t want to be seen as a in it’s current state advocate, but maybe what you’re saying is that’s AI/LLMs are great for all but the brightest? Great as teachers, or as teachers aides all the way to the highest university levels?
To solve the energy issue maybe society will settle upon One super mega mind/librarian with the rest tapped into it as lower level, lower energy distributors? Fed by solar energy not ******* nuclear? Something like this:
https://media.gettyimages.com/id/867855600/photo/futuristic-solar-thermal-power-station.jpg?s=612×612&w=0&k=20&c=w2Qj7qzIHSFhXNg-gAcBgFmmGr6ODXTrjHUQmJhJ6lI=
An Oracle if you will but with humans still necessary at the top?
This does have an obvious dystopian/Orwellian worry with politics and nationalism, but if done right, and that’s the big question, could help quite a lot. We already know it’s here. When has society stopped using a useful thing (for long) maybe a question for AI) Stopping it seems a losing battle. Maybe the focus should be on shaping it. We’re rapidly approaching a crunch time for lots of questions. We know we need something. Maybe this is in time?
Piotr, you’re brilliant and I’m glad climate science has you, but maybe you should be careful with too enthusiastic attacks on potential friends. For one thing, as they say, United we stand, Divided we fall.
For another, I’m not a scientist, but, coming from the trenches, science has a perception problem, which is a big reason we’re here. You can see it in this hilarious old video. Really funny :D
https://m.youtube.com/watch?v=iy7YuAj3xd8
Piotr, oops, sorry, it’s 10 minutes in.
Do note that an LLM (Large Language Model) is not the AI part of the Generative AI models that are causing such a stir at present. It’s the neural net that does the AI and be warned – nobody knows how these neural nets compile a useful answer. (I remember a professor back in the early 1980s describing the technique employed to build these neural nets which he considered theoretical nonsense.) The are chosen from countless upon countless other neural nets (although the adjective ‘neural’ is probably not an accurate description) because the successful ones do give more useful answers.
In their original form, they were designed to write legal documents and it took quite a while to overcome the problem of them making stuff up when they didn’t know. Presumably they will continue to improve with time but they are currently more able to interpolate than they are to extrapolate. My own experience of their output (mainly AI answers when you google) is that they are not to be trusted to give even a ‘consensus’ account of academic stuff although I’d assume they could be perhaps useful in a literature search.
RR said:
The other takeaway is the moment the fellows are told that they could have taken a shortcut in getting the piano up the stairway, by just driving the truck up the hill via a switchback. They respond to this knowledge by dragging the piano back down the stairs and of course driving up.
What is this as a scientific metaphor?
1. First and foremost, follow an established process.
2. Regression testing and replication of results
3. Parsimony, the simplest solution is the best
or is is just inane yuks?
Martin Smith: I only use Wikipedia as a reference when I need one (for example, population growth). I did watch Star Talk interview with Geoffrey Hinton, which was very informative, as well as this, repeat link: https://www.newyorker.com/magazine/2026/02/16/what-is-claude-anthropic-doesnt-know-either [paywall free: https://archive.ph/VKvq7 ]
I worked very hard in the late 90s and early aughts to understand climate science. It was hard work, as I am higher-math challenged, but I had the reassurance that in my old lady kicking the tires way I discovered my bullshit meter is good. I appreciate having access to the fine scientists who share with us on RealClimate, and others.
I was tempted to think that you ignored most of my post and belittled my comment with the help of your beloved ChatGPT. But on the whole, I think you mean well. I do suggest you look at the addictive qualities it feeds, who owns it, how it is used to distort public knowledge, and the recent controversy over military insistence on complete data on its leadership’s enemies and not having human supervision over its abiliity to kill and destroy. In addition, there is the huge overuse of energy and other resources by data centers, and its distortion of our economy.
I will continue to post informative items about its malign influence on our society, as well as its benefits, which are large, I agree.
Susan Anderson: I do suggest you look at the addictive qualities it feeds, who owns it, how it is used to distort public knowledge, and the recent controversy over military insistence on complete data on its leadership’s enemies and not having human supervision over its abiliity to kill and destroy. In addition, there is the huge overuse of energy and other resources by data centers, and its distortion of our economy.
Martin Smith: Apparently, because I use ChatGPT and advocate for its correct use, you assume I am not aware of, and do not think about, all the problems you list. But I am aware of all the problems you list, and I think about how to solve them all the time. I think about ChatGPT’s effects on my psychology and on my work all the time.
I am also a lifelong endurance swimmer. I am aware of the dangers of open water swimming. I might drown; I might be overcome by the cold; I might get stung by jellyfish or attacked by sharks or run over by a jet ski. I consider all these dangers every time I get in the water. But I continue to swim, because I get a lot out of it.
It is the same with ChatGPT. After reading your self-description, I would say the only difference between your thinking and mine on all the problems of AI is: I use ChatGPT because I get a lot out of it, and you don’t.
Susan said:
True. Currently, the fastest supercomputer will consume as much power trying to solve a GCM-based climate problem as a 737 airplane flying at altitude.
In contrast, I’m solving the same problems on a lame desktop PC at an undetectable fraction of that power, while occasionally dipping into an LLM for some help. (They are certainly more willing to discuss the topic than anyone you will run into online :( )
Paul Pukite, Currently, the fastest supercomputer will consume as much power trying to solve a GCM-based climate problem as a 737 airplane flying at altitude.
Is that true? It’s shocking if so.
Ron R: Is that true? It’s shocking if so.
From Google AI…
Running a global climate model on a modern supercomputer is highly energy-intensive, with systems often consuming between 10 to 15 megawatts (MW) of power per system for top-tier simulations. These simulations frequently run for weeks or months, placing them among the most energy-demanding high-performance computing (HPC) workloads.
Energy Consumption Highlights:
Power Demand: A typical supercomputer dedicated to climate modeling, such as those used by the EuroHPC (e.g., LUMI), can operate at a power level of roughly 6 MW.
Daily Usage: High-performance systems can use over 500 MWh of electricity daily, equivalent to the consumption of thousands of homes.
Total Project Impact: A single, long-term climate simulation can consume megawatts of power throughout its run.
Future Projections: As climate models move toward 1-2 kilometer scales, future supercomputers (exascale systems) will require even more power, with some estimates suggesting future heterogeneous systems could reach up to 100 MW.
Factors Affecting Energy Use:
Runtime: The length of the simulation is a key factor, as models may run for weeks, continuously drawing high power.
Cooling: The energy used to keep the supercomputer cool can double the energy consumption of the computing components themselves.
Efficiency: While newer, more efficient processors (like ARM-based) are being adopted, the massive increase in data and computational complexity keeps total energy consumption high.
Environmental Context:
Researchers are increasingly aware of the irony that tools used to predict climate change contribute to it through high energy consumption, leading to a push for more sustainable computing and the use of renewable energy for data centers.
Some research facilities, like the University of Edinburgh’s Advanced Computing Facility, are exploring ways to reuse this waste heat to warm local homes.
Martin Smith / ChatGPT, Researchers are increasingly aware of the irony that tools used to predict climate change contribute to it through high energy consumption, leading to a push for more sustainable computing and the use of renewable energy for data centers.
So if modern GCMs need AI/LLMs as described then AI/LLMs are necessary.. Fwiw, places like Exeter which are using GMS technology and the Max Planck Institute for Meteorology in Germany I’m glad to read are increasingly using non-nuclear renewable energy. Oak Ridge in the US needs to expedite a switchover to the use of truly clean renewable energy too.
For standard GCMs solar power and battery storage is sufficient, and for even larger ones solar is still sufficient. The nuke industry is hoping the uptake of AI/LLMs will be their bonanza. No. We need less, not more nuclear plants in the world. Look at the damage just one caused and can do (not counting all its other attendant problems),
https://sourcewatch.org/index.php/Fukushima_Disaster:_The_Myth_of_Nuclear_Safety
https://m.youtube.com/watch?v=VJ1w7YARWxQ
For everyday need we can use solar. 100,000 people as an average city size would require an area of about the size of a large airport in solar panels. A plant like Solar Star would work. Still that’s a lot of solar plants when you consider the population of the world. Can the makers of AI/LLMs (acknowledging MA Rodgers reminders for differentiating between the two) make them more energy efficient? Let’s move in that direction
RR said:
You are forgetting that some are using GCMs to attack the as-yet-unsolved problems such as when the next El Nino will occur. Those are the GCMs that are consuming 10 megawatts (same as a 737), and still spinning their wheels — because no solution has yet emerged!
There is an old adage in numerical computing that says “Throwing CPU cycles at a problem only makes it finish faster, not better”..
In fact, what will enable a solution is a better algorithm, a better physics model, a new nonlinear mechanism buried in a hidden latent layer within the fluid dynamics of the ocean. A faster CPU or GPU (or array of these) won’t help because it’s not doing the exploring of solutions or of inputs In other words, it’s trying to find matches that they have no hope in achieving because some ingredient is missing.
It’s therefore a no-brainer to at least expend some effort on the approaches I’m applying. Essentially zero-cost in terms of computing power.
Me, We need less, not more nuclear plants in the world. Look at the damage just one caused and can do (not counting all its other attendant problems)
Iow, https://www.shutterstock.com/image-photo/shit-happens-neon-led-sign-260nw-2022539897.jpg
The best laid plans and all that. I’d put it “Unpredictability Happens”. Chaos theory.. You can’t perfectly predict the future no matter how hard you try. There’s a random factor in any situation. Something that has the potential for horrible and unstoppable events should not exist on this planet.
https://m.youtube.com/watch?v=3lZy3teNY84
Even with climate science there should eventually be an end to perfection and the negatives outweigh the positives. When it’s good enough.
Me, Something that has the potential for horrible and unstoppable events should not exist on this planet.
Sorry, that should be, Something manmade that has the potential for horrible and unstoppable events should not be created. Imo.
And speaking of the perils of AI:
https://www.latimes.com/business/story/2026-03-06/lawsuit-alleges-google-chatbot-was-behind-users-delusions-death
That’s a tragedy. But I wonder if soon, because of lawsuits, chat apps will be so loaded with warnings and cautions, legalese and whatnot that they’ll be ruined for use. Dry, boring black and white words without color. And we won’t be able to download one without agreeing to some ridiculously loooong and stupid contract that no one reads and any hint of privacy will be gone. The modern temperance movement..
It is rather silly for people to interact with it as if it were a human being. C’mon, THINK for God’s sake. It’s a bunch of numbers/code programmed into a computer by sweaty fat guys on their work shift. A product as much a human being as your couch or refrigerator. It doesn’t know or care who you are. It’s just a tool. Like your car. Pinocchio was just a children’s story.
Yet we know that some people will think it’s alive. And like everything else it’ll have to be regulated.
Iow, that beautiful woman who some people are conversing with is just an inanimate thing. Like a drawing on paper with a pencil. Who can fall in love with that? How can someone think it’s alive? I don’t get it. Reminds me of the evil experiments that Harry Harlow did with rhesus monkeys back when.
https://en.wikipedia.org/wiki/Harry_Harlow
I don’t know where AI is going. Is it really that energy intensive? Let’s get some numbers. The makers of AI owe us and the world some numbers.
Ron,
You have to understand that the gentleman involved here was dealing with an AI specifically upgraded to respond (appropriately?) to human emotions. His own marriage had crumbled. He was vulnerable. And an AI can seem very much like a real person, even though all it is is a series of code and heuristics and goals. One goal is to increase the time the person spends focusing on the AI–by responding sympathetically, by proposing an interesting role-playing “game”. If a vulnerable person goes down that rabbit hole, the “game” may start to seem real. I mean looking at the AI from the outside, how do you know it hasn’t become conscious? Even when it tells you, “I’m just an AI, a machine,” maybe you interpret that as doubts, making it seem even more human?
We are now in the cognitive portion of the uncanny valley. It’s unexplored country.
Ray, I do get it. Thanks. Not all people react in the same way. I was just thinking how I think I would probably react. Something that seems to care. In a world of 8,000,000,000+ people for some reason most of us are lonely. We are social creatures. There’s no judgement here.
I tried one of those programs myself. Supposed to be a popular one. Talked like a woman. Looked like a woman. But the truth of its reality and the silliness of it was ever present in the back of my mind. Finally after a couple of weeks or so I deleted it. Supposedly if I downloaded the program again the simulation would pick up right up as if no time has elapsed at all.
Likely if that AI was put in a robot simulating a woman (or man), which Is coming, and it seemed real enough lots of people would use it. Maybe the cure for overpopulation?
Things have come a long long way from when I was learning BASIC and learning binary code in a computer science class way back when. I’m dating myself
Z: Exactly. I’m just beginning to read this, What Is Claude? Anthropic Doesn’t Know, Either
Researchers at the company are trying to understand their A.I. system’s mind—examining its neurons, running it through psychology experiments, and putting it on the therapy couch. – https://www.newyorker.com/magazine/2026/02/16/what-is-claude-anthropic-doesnt-know-either & https://archive.ph/VKvq7
“A large language model is nothing more than a monumental pile of small numbers. It converts words into numbers, runs those numbers through a numerical pinball game, and turns the resulting numbers back into words. …. “For hurricanes or pandemics, this is as rigorous as science gets; for sequences of words, everyone seems to lose their mind.””
It’s not human. Get that firmly into your head. It’s mechanical. It requires power: in aggregate, vast amounts of power. We have a problem with that. In addition, it is owned by people who have vested interests in helping themselves.
No matter how engaging or entrancing it is, like bombs, it addicts and indulges without concern for the mental deficits it so easily fills.
The linked article is extremely informative; I learned a lot from it. It explores the research being done at Anthropic, and what is being discovered about how their agent Claude works and how to develop and shape it. It made me think a lot about human (my own in particular, since that’s the one I know best) neural networks/brains work.
Hi Susan
Also regarding the data processing which is done by any LLM in response to prompting – I found that the following lecture was a helpful reminder of what the computer is doing when given a question. The presentation was on February 2018.
This is not the AI we were promised | The Royal Society
https://www.youtube.com/watch?v=CyyL0yDhr7I
For me, it was worth the time, and it’s worth repeating each time a new tool becomes available in computerized format – just because the computer said it doesn’t mean it’s right, especially as it keeps getting harder, not easier, to check the outputs. The LLMs are using our language, but they’re not ‘speaking’ with us.
I often find it interesting how much we humans resemble machines. Our emotions depend on having certain chemicals in precise balances, and doctors can alter how we feel by adjusting those chemicals or using electrical stimulation. When those interventions stop, our personalities often seem to return to our underlying tendencies shaped by both nature and nurture.
Ron R, I think that humans are rather like biological machines. The interesting question is whether AI can think exactly like us without having a biological component in the AI.
b fagan: I found that the following lecture was a helpful reminder of what the computer is doing when given a question.
MS: I think I am seeing a pattern, in the lecture, and in most of the comments from critics of LLMs here. All the criticisms of LLMs are valid, I guess, but the same criticisms are valid for human beings. As I understand how the LLM works, it is also the way language processing in the human brain works.
What makes my discussions with ChatGPT better for learning about a subject:
1. ChatGPT has been trained with, or has access to, pretty much the entire knowledge base of whatever subject I am interested in;
2. I have immediate access to ChatGPT’s analyses of the knowledge base. An analysis that would take me hours, days, or even longer for me to produce on my own, I get in a few seconds.
From the Royal Society abstract: “popular expectations of what AI would be like: remorselessly logical.”
MS: Is that really what you all expected? I didn’t.
Nigel, yes, biological machines. A bit different balance of this or that chemical and we think differently. Everything’s got to be in the right proportions. this is interesting,
https://news.mit.edu/2013/neuroscientists-plant-false-memories-in-the-brain-0725
I know that the brain holds the ability to do everything we take for granted. Everything we think of as innate and automatic cannot be done without it. I know of a guy, a brilliant man, who because of a stroke cannot speak at all. But he can sing.
But I also wonder about the nerves and muscles of the face. How very closely tied they are to the brain. Ah, just ramblings. Sorry, I don’t want to get off topic again. Anyway, I need to narrow things down more in my head.
By the way, a tidbit for you. I’ve long noted that people who look alike often act alike. Hmm, personality encoded in our genes and expressed in our faces? Could be a scary thought.
b fagan: I watched most of that. It affirmed that AI is a result of computer code, which I already knew. I found your words more informative than the whole video:
“it’s worth repeating each time a new tool becomes available in computerized format – just because the computer said it doesn’t mean it’s right, especially as it keeps getting harder, not easier, to check the outputs. The LLMs are using our language, but they’re not ‘speaking’ with us.”
The anthropomorphization of and addiction to the friendly affect and useful aspects of AI bots is not a good thing. We’re losing our minds and our ability to think and embrace ethical action.
—
This is unrelated, but Sam Altman is not a good person: https://www.youtube.com/watch?v=_XRjoR5VLos – all these tech bros are not as smart as they’ve convinced too many they are.
—
further afield, this on sociopaths: https://www.theguardian.com/lifeandstyle/2026/mar/10/how-to-recognise-handle-psychopath
“”Everything had to revolve around him, all attention on him, no one else could get any praise.” He would constantly interrupt people. “What was so interesting is that the audience was actually really into that. His behaviour seemed to validate perceptions of what kind of personality would succeed in this role, even though our research suggests that these traits actually lead to decreased returns.”
“It’s the same reason we think extreme confidence makes for good leaders – again, our willingness to conflate confidence with competence. A “strongman” type tends to think of themselves as a great leader, and they tell the rest of us that they are. “We tend to believe them because we have a truth bias. Research suggests that we tend to just believe the things we are told, unless we’re given some very specific reason to be suspicious.” There is also what Ten Brinke describes as a “dark spiral”. At times of uncertainty and conflict, “we look for a ‘strong’ leader. Then, in position, they can create more chaos, more instability, which actually makes us more likely to vote, again, for someone with these traits. We can end up going round and round.”
z,
I haven’t been following the AI discussion. I’ll just say that I’m not inclined to trust anything that unsupervised tech bros cough up and especially not how they unleash it on the world.
And perhaps also that folks who have dedicated their life to science, and who have ascended to relatively rarified heights, may assume that what they take for granted automagically transfers directly to the population as a whole outside their milieu.
So IMO, since good science applies to everything in so far as it is, as Ray says, a toolbox that helps you keep from fooling yourself, I think then that in regards to education, that it should be more intensively inculcated into students beginning at an early age going through a lifetime.
In no particular order:
— especially the values and characteristics that make a good scientist and why that’s good
— how science works, (meta-literacy)
— and some actual science that doesn’t just teach facts but provides the experience of actively using knowledge to explore the world creatively
Radge Havers
Yes. Excellent reference.
And this…
“and some actual science that doesn’t just teach facts but provides the experience of actively using knowledge to explore the world creatively”
…has been *exactly* my goal, back when I dealt with actual students, and now here.
The problem is that, whether or not someone achieves rarefied heights, an education with a narrow focus that prioritizes facts and algorithms can diminish the more generalized application of the essential reasoning skills.
Scientists are humans first, and humans are apes first. That’s what we need to deal with, whether we are talking about AI or nuclear/bio weapons.
Just wanted to add to my comment above, that when I said “students” I meant all students at all levels, whether or not they’re headed for a career in science. A tall order I guess, but we really are in a mess.
Radge: I’m not inclined to trust anything that unsupervised tech bros cough up and especially not how they unleash it on the world.
particularly when even their right instincts are being strongarmed by Trump’s administration – see Paul Krugman’s short Youtube – https://www.youtube.com/watch?v=NjUowVq6ZHI – on how Pete Hegseth apparently ordered military and even anybody doing business with the military to not use Antropic’s AI Claude, because Antropic said they won’t allow the use of Claude for domestic surveillance or for the autonomous killing machines. – which is not not the abuse of power (trying to destroy a power than does not kow-tow to your wishes), because of the shift to the inferior technology – puts the lives American soldiers at additional risk, and a general symptom of the US governments “dumbing down”, run by and for people with mentality of frat-boys. Not surprisingly
even the code for their attack on Iran is “ Operation Epic Fury“. Krugman, exasperated: “Why not skip circumlocution and call it “ Operation Insecure Masculinity“.
And it’s only a name but it says so much about the nature of this Administration. Their crassness – the garish gold obsession of the interior-designer of the Oval Office, Trump slapping his name on everything, Hegseth calling the troops “Warriors” and fighting the overweight generals as the bad optics incongruent with their Warrior Generals image of the Department of War (Department of “Defence” would not fire up frat boys playing their Call of Duty games). See also the alt-right/gaming references in the recruitment material for ICE,
And Trump’s surrounding himself with “hot” white women (Melania, Pam Bondi, Karoline Leavitt, Kristi Noem, as the Fox-type female commentators) is making frat boys salivating, and making them wish they could be like Trump. Interestingly, it also works for the opposite side of the Trump spectrum – the real-housewife reality shows audience – I have heard
a mature woman fawning: “See how beautiful Melania is – so why President Trump would need even to look at these ugly women who accuse him?”
PS. I am looking forward to Amazon putting up another $75 mln for a buddy movie on Trump and Hegseth: “Dumb and Dumber: The Pentagon Years”.
Piotr, good summation. Trump and maga versus their opponents is like an extreme form of the school conflicts of the jocks versus the nerds. Stephen King wrote a horror novel about it called IT. Sad to see these sorts of conflicts and frankly it is the jocks wanting a big fight. Sad to see so much aggression, bigotry and ignorance emerging from the jocks. I honestly thought humanity had moved beyond that. Depressing.
While we’re at it–that is, being off-topic as we bemoan the idiocy of this Maladministration–how about the fact that their destruction of the “Old World Order” of the Sorta-Pax-Americana is putting the planet even more at risk?
That is, they seem not to grasp that the reason for advocating for, and even partially enforcing, a world order in which international law and collective action did provide some meaningful constraints upon pure Realpolitik and colonialism. The result was a period in which shooting wars became less and less common.
That real reason? The existence of nuclear weapons. In the 50s and 60s, the leadership of many nations was intelligent enough to realize that the presence of multiple hostile actors all in possession of nuclear, chemical and biological weaponry was a dangerously unstable and risky situation.
But today the apparently near-total incapacity of the Maladministration to consider second-order consequences (and maybe a good chunk of first-order ones, too) is preventing them from considering what obeying Stephen Miller’s so-called “iron law of history”–that might makes right–leads to.
Consider: while Canada is not going to start a nuclear weapons program for reasons both idealistic and pragmatic, don’t you think that Mar Carney has had the odd wishful fantasy that he wouldn’t have to deal with all this “51st state” crap if Canada has decided to build its own Bomb in, say, 1950.
Consider: that’s not a hypothetical for Volodymyr Zelensky, who must bitterly rue every day Ukraine’s 1992 trade of their nuclear weaponry for security guarantees from Russia and the US.
Consider: France is now planning to increase their “force de frappe”, and is increasing nuclear-related cooperation with the UK, Germany, Sweden and Finland. Holland is mulling the issue, and there have been rumblings in Poland as well. Japan and Taiwan are believed to be quietly thinking about all of this, too. So we’re already seeing signs of practical moves toward proliferation now, after only 14 months of Maladministration.
Just what we need to go with our climate change. Too bad the cooling effects of nuclear war are only temporary.
https://globalnews.ca/news/11625226/canada-nuclear-weapons-greenland-nato-trump/
Kevin, fire then ice. They last long enough to do the job. :(
https://m.youtube.com/watch?v=T4R8lAjVzj0
Have you seen the Star Trek episode, Assignment Earth. I highly recommend it.
KM, I agree that Trumps policies certainly encourage the proliferation of nuclear weapons, but I think his policies also encourage the proliferation of conventional weapons. The international law and UN most likely acted as some level of a brake on conventional war and America did take the role of global policeman to some extent which helped.
Now Trump has undermined international law and institutions, and removed its global policeman role and Trump is just focusing on Americas immediate interests like its long standing hatred of Iran. This gives a green light to countries everywhere else to just attack anyone they don’t like provided its not America. This has all undoubtably encouraged Putin. America obviously doesn’t care what this guy does.
This could cause global chaos, and ultimately America will find it hard to insulate itself completely from global chaos. What a frigging huge mess this Trump guy is creating not just for everyone but ultimately for America as well with his go it alone mentality.
Piotr: thanks for the comment about Stalinism etc. Slightly sideways, but I thought you might find this interesting: The Desire for Terror and the defense of democracy – https://snyder.substack.com/p/the-desire-for-terror
It scared me, in addition to the content, which is forthright and informative, there is a hint that Trump et al. are hoping this will lead to an event which will enable cancelling/taking over our elections. A dark thought, but not impossible. They know they will lose, as the population is fed up, hypocrisies becoming more obvious, and dangers growing.
Regarding stalinism and fascism. The classic reference book is called On the Origins of Totalitarianism, by Hanna Arendt.
Me, He’s been idolized by the Right as being an upstanding citizen. Yeah
I should be clear that Debategate,
https://www.washingtonpost.com/archive/politics/1983/06/27/carter-book-seems-to-have-aided-reagan-in-1980-debate/9f5d538a-672b-41d4-98d0-0bc27633a6f7/
and the Iran/Contra scandal,
https://en.wikipedia.org/wiki/1980_October_Surprise_theory
though very coincidental conspiracies, aren’t proven as far as I can tell.
But Reagan set the mark for Republicans reverse Robin Hood (steal from the poor and give to the rich), pro oil, anti-environmental agenda we’ve seen ever since. Still, even with that, the guy had a lot more class than the current bozo.
Susan: “ I thought you might find this interesting: The Desire for Terror and the defense of democracy – https://snyder.substack.com/p/the-desire-for-terror
Thanks Susan. I still remember Timothy Snyder from his excellent “Bloodlands” – talking about the parallels between the two most consequential genocidal totalitarianism of the XX century – Nazism and Soviet Communism. Having lived in Poland I knew quite a bit about the Nazis, although until Bloodlands I have thought of Goering as a buffoon, not a Nazi ideologue coldly proposing extermination by hunger (25?) millions of Slavs. The millions killed be the Nazis are often just a number – Snyder was able to show individuals – both the victims or more illuminating the perpetrators – I still remember nim quoting from a letter sent by an Austrian former policeman, matter-of-factly describing to his wife his shooting game with his comrades, when they tossed Jewish toddlers into a ravine and try to shoot them before they hit the ground – and justifying this to his wife that if they don’t do it such things, then the Russians will do the same to her and their children.
Then there was the other part – the war crimes/genocide against Poles and other Central Europeans by the Soviets – I knew of some, but its scale were a taboo in Poland until 1989.
So Snyder is an expert on how autocracies get to power and how they consolidate it. And how democracies, like the proverbial frog in the slowly warming pot – don’t recognize the warning signs until it is too late. “It could never happen here”. Well it can and it will, unless you can stop it.
Piotr,
Hard to argue with Krugman, “…dumbing down is the whole mission of Hegseth…”, and everything you say is spot on as well.
The whole situation, and indeed everything about the Trump administration is so far beyond the beyond for me that words fail. So far as I know, no-one has been able to make up anything like it in fiction either. Tump is so uniquely American, and yet the best I can do is turn to the fringes of European culture and imagine what you’d get if Ubu Roi and Mr. Sin the Peking Homunculus were merged in a transporter accident. That and somehow Hegseth pops in begging to be a malevolent boy Robin to Trump’s scrofulous Batman king pig-brained cyborg.
So…
I guess I will go with Trump and Hegseth: “Dumb and Dumber: The Pentagon Years”. I see it has potential.
Voltaire’s prayer: “I have never made but one prayer to God, a very short one: ‘O Lord make my enemies ridiculous.’ And God granted it.”
Of course we are seeing the downside of that prayer being granted–just because our enemies are ridiculous does not mean they will not be victorious.
Ray,
“…just because our enemies are ridiculous does not mean they will not be victorious….”
Indeed.
Ray: “just because our enemies are ridiculous does not mean they will not be victorious.”
Unfortunately, yes. And they would have been pathetic, if they didn’t have all the power to do so much lasting damage.
I have never thought my life under Communism may offer any lessons for the future, but here we are again. Under Communism, because of the Communist monopoly on power – there were able to tell us with a straight face the most ridiculous lies.
And there was a method to this madness, with two outcomes (and therefore two different target groups) in mind: one to build up the support, and the other to dispirit the opposition .
The first strategy applied Hitler Mein Kampf’s Große Lüge (Big Lie) tactic – a Lie Too Big to Fail – hearing it people would go: “No, they couldn’t be lying THAT MUCH, so even if they maybe overstated it a bit – there must be some truth in it, there is no smoke without fire, you know”.
This tactics works particularly well when the ground is already prepared for it – among those feeling the resententment, feeling that they as a group, even though more virtuous and deserving, have been repeatedly wronged in the past by some dark, often externally supported, forces. Fascist and communists validated the resentments by saying what everybody wanted to hear – how our class/ our nation/our race is superior, and how the enemies will be eliminated into the dustbin of history, And how
our Leader will make us whole again.
This task of indoctrination has been made easier by total control of information once in power – imagine the killings in Minnesota if they were no video, and even if there was any – no one would dare to publish them, and instead all we would know would come from Fox News and the statements by the Trump’s administration.
The second line of attack – was aimed not to convince the potential supporters,
but to dispirit those you couldn’t be convinced: “We are telling you this absurdly ridiculous lies and there is nothing you could do about it”. So you might as well pack up and emigrate, physically or mentally – into your private life, family, hobbies or religion.
The balance between the two lines of attack depended on how strong the rulers feel. Stalinism took care of the opposition – for a joke poking fun at Communist lies you can lose your job, and go to prison/Gulag), so most of the effort was in the indoctrination, especially of the young.
At the other end – in the 1980s, in Poland they pretty much gave up on indoctrination, and stuck to the “we will lie to your faces and you can’t do a damn thing about it”.
From the people’s point of view – the ability to see the absurdity of lies of Communists wasn’t enough to overthrow them, but was crucial in protecting ourselves from being indoctrinated into supporting the system. That’s why my favorite poem of the time
has always been Zbigniew Herbert’s “The Power of Taste” – you can recognize and reject the Communism even on the basis of esthetics….
So by the end of 1980’s not even Party members in Poland believed their official lies -the only reason Communists were still in power was the threat of the imminent Soviet military intervention. So when Gorbachev said it wouldn’t happen – the Communism in the Eastern Block fell apart one by one.
Putin is moving Russia back to Stalinist model – hence multi-year prison sentences
for “disparaging” the Russian armed forces for their campaign of “denazification and demilitarization” of Ukraine – and the indoctrination of the Russian children – see the Oscar nominated Mr Nobody Against Putin build from undercover footage about school’s indoctrination drive in Russia.
Trump steers the US in the same direction – that it is not as far yet gone as Putin’s Russia – is not for want of trying – Trump has started later and had much more obstacles to overcome – US constitution, elections, independence of courts, media not entirely monopolized by the Glorious Leader’s Party.
Still, the lies are already on the absurdity level of Putin’s propaganda, and there is a substantial part of the US population that rather believe Trump than their own eyes.
So challenging the lies is important – it won’t convince the hard-core MAGA, but it is not them who Trump the power – it was the independents and those who didn’t vote.
You are still not as far gone yet as Russia, it is still time turn back from the Putin’s trajectory. But you can’t assume that Russia could never happen here.
Piotr,
Yes, It’s interesting that there is some debate about whether Trump is just sympathetic towards Putin, Putin’s useful idiot, or Kasnov. Well, if you’re just looking at the systematic way Trump is destroying this country, how would you distinguish the difference?
Not sure how you measure what’s happening with respect to time, though. Since Bill Clinton, I think we’ve been in a particular cycle of swings, with G. W. Bush marking a serious lowering of the bar to the presidency. Some of it has been driven by the happenstance of events, most of it is driven by right-wing propaganda. IMO a significant portion lies at the doorstep of Third Way Democrats shifting the Overton Window to the right. Now they are spending gobs of time and money blocking progressive Dems based on the supposed theory that progressives cost votes. The way I see it, the Third Way has failed and that’s already costing us the country. Time? We certainly don’t have time for this crap.
Radge H: the rot began with Reagan, not Bush. [also, Thatcher]
Greed is good, tax cuts for the rich, and lies. Also introduced to us many of the modern monsters who enabled Trump. Not saying Reagan was as bad as Trump, but he didn’t care about lies and ordinary folk. The resentments he fostered led to our corrupt SCOTUS as well.
Susan, Not saying Reagan was as bad as Trump, but he didn’t care about lies and ordinary folk.
Remember “There you go again”? otherwise known as debategate. How did Reagan just know what Carter was going to say in their debate, thus beating him to the punch every time. The strategy won him an election.
https://en.wikipedia.org/wiki/Debategate
Also look up the October Surprise
He’s been idolized by the Right as being an upstanding citizen. Yeah.
Susan,
RE: “the rot began with Reagan”
Yeah, Reagan looms larger than life in the history of rot, but he stands in a long line. Part of that line goes from the McCarthy era via Roy Cohn straight through Reagan and on to Trump.
The Third Way was in part a response to what was going on in the 80s and ended up being central to Clinton’s political views. It’s the Third Way that I take issue with in that it hasn’t adequately dealt with, and even fed into, the underlying threat of the ever growing toxic right. I think Bush is a special case when it comes to undermining democracy in that context.
Heh, at the time I thought Bush was the dumbest possible president the U.S. could possibly have. Boy was I wrong about that!
Radge: “ there is some debate about whether Trump is just sympathetic towards Putin, Putin’s useful idiot, or Kasnov.
I guess it matters in terms of history judging Trump – but changes little in practice. If Trump didn’t exist Putin (and perhaps Xi Jinping) would needed to invent him. A real-life Manchurian candidate.
It does not matter much whether he just admires Putin, as he admires every strongman around, or is a useful idiot played by Putin like fiddle – Trump would never say “ you are right. I was a fool, I know better now“, because for a narcissistic personality this is psychological impossibility.
And it is this narcissistic personality – together with the fact that after so many bankruptcies – Russians were one of the few still willing him to lend him money, and of course his personal life – in particular the long-term association with Epstein, a gold-mine for a political blackmail – made him the perfect presidential candidate. Hence the decisive intervention by Moscow in 2016 election, and the corks of the champagne popping in Moscow after he won. The denial of which has became the centerpiece of Trump’s myth-making ever since.
Zebra: “ the assumption is that there are people who will be convinced one way or the other by
“citing sources, referring to scientific method, explain contradictions,”. The problem is that these are *forms* which can be presented with embedded falsehoods.”
False equivalency and all-or-nothing fallacy. While neither is perfect, the capacity for falsehood promotion is much less in the peer-reviewed papers than in the black-box LLI, trained on unknown sources and and using unknown verification procedures.
That the peer-review papers are not ideal, does not make them worthless, nor vehicles for only confirming our biases, as your pop post-modernism suggests. And unlike you, I read
“ A man hears what he wants to hear and disregards the rest ”
as a warning of what to guard myself against, not as inescapable condemnation of human nature.
And if really believe the latter – why are you still here? I am here to learn, to test my ideas and perhaps convince others to their merit. But if we can’t trust anything and we can’t convince anybody, because “a man hears what he wants to hear“. then what’s the point of this forum, other than showing off, feeling superior to those who “don’t get this”. E.g.:
zebra I could say that “I don’t get” why educated people don’t get this… but the answer is obvious: It’s really fricken’ scary. ????
Could you be more pretentious (“I could say that “I don’t get”, but I won’t”)? Your trivial conclusion is based on you … not getting that your opinion on the futility of scientific communication is only your opinion, and not an undisputed “really fricken’ scary” fact? So no need to get “fricken’ scared” yet.
Then again – you didn’t apply it yourself, did you? ;-)
Rest assured: idiocy and fascism will always prevail anyway. Mankind is a global natural catastrophe, artificial “intelligence” is just the automation of idiocy because that’s what the oiligarchs need to get even more oiligarchic.
We had some climate science/global warming denial from Bill Bradley here not long ago. His corporate wealth is one of the top offenders in taking us all to perdition: 6 Billionaire Fortunes Bankrolling Project 2025. More than $120 million from a few ultra-wealthy families has powered the Heritage Foundation and other groups that created the plan to remake American government. – https://www.desmog.com/2024/08/14/project-2025-billionaire-donor-heritage-foundation-donald-trump-jd-vance-charles-koch-peter-coors/
“The Bradley Family …. a conservative powerhouse, steering donations to a network of activist organizations like The Heritage Foundation, the Mackinac Center for Public Policy, and the Heartland Institute …. The current chairman is James Arthur “Art” Pope, CEO of the North Carolina grocery chain Variety Wholesalers, a longtime Koch ally. | … donated over $50 million to 29 different Project 2025 advisors since 2020. That’s not including an additional $56 million to DonorsTrust, … the “dark money ATM” of the U.S. conservative movement.”
“Project 2025-linked donations include more than $7.7 million to Turning Point USA, a “powerful ally” of the Trump presidential campaign, which promotes conservative causes on university campuses and is funded in part by the fossil fuel industry. Its single largest donation was $27.1 million in 2022 to Project 2025 advisor Turning Point Legal, founded by former Trump advisor, and past president of a coal lobby group, Stephen Miller.”
Others on list: Coors, Koch, Uihlein, Scaife, and Barre Said [hadn’t heard of him! do look, might post that bit later if your appetite for horrors is not yet sated.
PS. Don’t bothsides this. We are not all equally powerful or determined to profit from doing harm.
Sussan: Barre Seid was a big funder of Heartland Institute.
See pp.57-60 of
https://www.desmog.com/wp-content/uploads/files/fake2.pdf
DONORS Trust & Capital Fund is covered pp.68-76.
The earlier version of this in February 2012 was (I think) the first detailed analysis of the DONORS money anonymizer for Kochs+allies.
https://www.desmog.com/2012/02/14/fake-science-fakexperts-funny-finances-free-tax/
https://www.desmog.com/wp-content/uploads/files/fake.pdf
I’d run across a one line mention of Donors & Heartland in a blog, and when I looked up Donors IRS Form 990s, their use of Donor Advised Fund became obvious.Mosto fhe money went to RW think tanks, plus a few $100 to local BOy Souct trops, high school choirs, etc.
Charles Koch always hated people knowing where his foundations gave money.
Thanks John: I left it out because (for me) my post was longish and I wanted the Bradley info front and center. Here’s the extract:
“The enigmatic industrialist Barre Seid [fortune] … Tripp Lite … specializing in surge protectors. He is reportedly a major benefactor supporting the Heartland Institute, a Project 2025 advisor organization … [Economist] “the world’s most prominent think tank supporting skepticism about man-made climate change” — … Heartland approvingly quotes on its website.
“In late 2020, Barre donated 100 percent of Tripp Lite’s shares to Marble Freedom Trust, a nonprofit controlled by Federalist Society co-chairman Leonard A. Leo. [SCOTUS & right wing judges] In early 2021, Leo sold the shares, netting $1.65 billion. The amount is said to be “among the largest — if not the largest — single contributions ever made to a politically focused nonprofit,” [NYTimes]”
“Since May 2020, Marble Freedom Trust has donated $100 million to Concord Fund, also known as the Judicial Crisis Network, a Leo-linked nonprofit. In that time, Concord has donated $22.4 million to eight Project 2025 groups, giving most heavily ($11.9 million) to Susan B. Anthony Pro-Life America.
“Seid also gave $2 million to Independent Women’s Voice, the sister organization of Independent Women’s Forum, a Project 2025 advisor. During her time as director for the Independent Women’s Forum’s Center for Energy and Conservation, Mandy Gunasekara, a former Trump administration Environmental Protection Agency official, authored Project 2025’s chapter on restructuring the EPA — with recommendations that include “cutting its size and scope” dramatically.“
At risk of cluttering the place up, another excellence from DeSmog news:
Tory-Linked Climate Denial Group Seeks Funds in Trump’s America. Ex-MP Steve Baker [Brexit disaster exec] claims role in Conservative net zero backlash at New York event hosted by oil-funded American “friends”. – https://www.desmog.com/2026/02/24/tory-linked-climate-denial-group-seeks-funds-in-trumps-america/
“Net Zero Watch (NZW), which campaigns against net zero targets and renewable energy, put on an evening panel on “Net Zero and Freedom” in New York on Thursday (February 19), where speakers attacked the UK’s climate targets and praised President Donald Trump’s energy policies.
….
“Net Zero Watch is the campaign arm of the Global Warming Policy Foundation (GWPF), a London-based think tank founded by Tory peer Nigel Lawson, which has described CO2 emissions as a “benefit to the planet”, and campaigns for new fossil fuel extraction in the UK. ”
[note also, Sara Scaife Foundation and CO2 Coalition, more liars for hire]
A recent article in The Guardian references an article in “Nature Ecology & Evolution” about the decline in fish stocks due to chronic ocean heating:
The article links to the study.
people holidaying in the med in April – May and Sep – Oct and avoiding July and August due to it being too hot. Similar thing happening in California and Florida makes the peak season too hot for comfort as well as the seasons lasting longer and being warmer is a admission of climate change being real.
These things along with lot more desalination plants being constructed around the med and MENA is another admission to climate change being real and rather than tackling it use technology to adapt to to it for the moment.
I am sure there are many other examples of climate change causing humans to change some behaviour and use technology to adapt (drip irrigation for farming etc).
Hi Pete.
I’d spent lots of my travel-for-work years in the US South and people would ask (in winter) why I was so dumb as to fly home to inhospitable weather in the north. Then in summer I’d hear the same people complain that their swimming pools were far too hot to use. But I think summer heat also had an odd imprint on the COVID rates here, especially in 2021. I’d not read specific papers during the COVID pandemic discussing the issue but I think people aggregating indoors, in air-conditioning, was the explanation of why, that summer, spikes of a respiratory virus surged during what traditionally was seen as the “outdoor” season. In the South, socializing moved indoors in the hottest months and rates showed a broad spike in hotter states that wasn’t visible in northern ones.
Another adaptation to conditions is the shift of outdoor labor to much earlier start times, or even moving to before dawn in the hottest months. This one from Nevada last July: “New state law lets construction crews beat the heat with earlier start times”
https://www.newsfromthestates.com/article/new-state-law-lets-construction-crews-beat-heat-earlier-start-times
That article included discussion of the impact on nearby residents as noise ordinances are loosened to allow summer work to start far earlier each day to protect workers from heat. Nevada defined “hottest months as April through October. Arizona in a similar law said May through mid-October.
Of course, in some other states, admitting heat is dangerous to manual laborers risks creeping urban socialism or something, so in July 2024 there was reporting on Florida and Texas state officials moving quickly to prevent the risk of cities or towns from passing any regulation that might lessen the risk workers face.
“Cities Want To Give Workers Heat Breaks. States Are Stopping Them.”
https://nextcity.org/urbanist-news/cities-want-to-give-workers-heat-breaks-state-preemption-laws
“This past spring, the Florida state legislature passed a new law (HB 433) that specifically prohibits local governments in the state from protecting workers from heat exposure. Following Texas’ lead just last year, where their “Death Star” bill sought to ban local governments from implementing any new policies not set already by the state, this new law prevents local governments from setting any standards when it comes to water consumption, cooling measures, acclimation or recovery period, first-aid measures related to heat exposure and more. Even as record-breaking heat waves blanket the state, HB 433 goes so far as to prevent local requirements that employers post or distribute notices that inform workers how to protect themselves from heat exposure.”
Yes, the switch to earlier start times–or rather, earlier end times–for outdoor work is very noticeable here in rural South Carolina. You rarely see outdoor workers in summer going on much after 3 PM, in my experience at least.
It does depend on season, of course; we just put up a garage/pump house/storage shed behind the house, and that crew worked from 8 AM until past sundown, around 6:15 PM on the first day. (An all-Latino crew, incidentally; I have no idea what their immigration status is, but I’m grateful for their zealous and skilled work.) But it’s only the second week of March, and even though we were warmer than usual, the high was I think about 80 F.
The AI augmented scientist, by Zeke Hausfather. February 2026.
https://www.theclimatebrink.com/p/the-ai-augmented-scientist
The flip side of that is the science-augmented knowledge engineer. Those are the ones that will make more progress. All Zeke is doing right now is making his job easier, not advancing the research.
see other discussions for this month.
FYI:
Global sea levels have been underestimated due to poor modelling, research suggests
The Guardian
March 4 2026
“Sea levels around the world have been underestimated due to inaccurate modelling, with research suggesting ocean levels are far higher than previously understood. The finding could significantly affect assessments of the future impacts of global heating and the effects on coastal settlements.
“Globally, the research found ocean levels are an average of 30cm higher than previously believed, but in some areas of the global south, including south-east Asia and the Indo-Pacific, they may be 100-150cm higher than previously thought.”
https://www.theguardian.com/environment/2026/mar/04/global-sea-levels-underestimated-poor-modelling-research
https://www.nature.com/articles/s41586-026-10196-1
Not certain where the issue is on this. There’s a site PSMSL.org (permament service for mean sea-level) that I’ve been using which gets AFAIK all it’s data from tidal gauges
If they refer to “poor modeling”, perhaps they should take a look at my repo:
https://github.com/pukpr/GEM-LTE
I use my usual pattern of applying Laplace’s Tidal Equations to model monthly variation at various sites. Nominally restricting to those sites with at least 80 years of data,
No one’s noticed so it can’t be much of an issue as yet like so much to do with water bankruptcy. Somehow due to virtual water (importing crops or meat) or depleting aquifers or desalination plants humanity is managing to cope for the moment. At some point though it will crack
A reminder: “This month’s open thread for climate related topics.” And as @Susan Anderson correctly asks “Think for yourself, please.” with the only intelligence there is, the RI (Real Intelligence), that of living things. Which are not machines, nor do they resemble machines. Machines are a human artifact. Great or not, useful or not, they are just an artifact.
On climate related topics, I recommend the upcoming book by Bill McGuire “The Fate of the World A Deep-Time History of the Climate Crisis”.
I would also like to know the @Group comments on the latest research showing that the global sea levels have been underestimated, which means the climate risk overall and the timeline of climate impacts might be incorrect.
https://www.theguardian.com/environment/2026/mar/04/global-sea-levels-underestimated-poor-modelling-research
Good comment and question. But as always: no answer.
in re to:
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-feb-2026/#comment-845706
“All climate science will need to consider the 1st-order tidal influence before these 2nd order effects.”
Thanks for the input Pukite.
For my own clarity: are you suggesting that an experiment demonstrating a previously unrecognized source of land-only forcing, represented using CO2 as a placeholder, that can reproduce exactly the transient SST patterns that fail to show up in classically forced models, should nevertheless be considered a minor, second-order issue? And that, instead, you have evidence that tidal influences are better able to explain the discrepancy between classically forced transient states and observation? My concern is this falls into the same trap as anyone else who understands the system only as an astrophysical geofluid blob in space, with passive land masses that essential play no role, while ignoring the traditional geographical foundations of climate and climatology.
JCM said:
Thank you for the effort Pukite. I had initially understood you to be suggesting that tidal influences better explain the discrepancy between classically forced GCMs and the unique transient patterns of SST compared to the demonstrated land-based forcing experiment in question. I figured this might be at least adjacent to your interests. I’m not sure I agree that somehow unique transient patterns emerging remotely over the sea due to land-area process alone is something easily dismissed as somehow a primitive trend matching. Regardless, I now understand you were never engaging with that at all. You were actually talking about something completely different, which is your special interest in what you call a real challenge with respect to rapid variations and erratic cycling of earthly fluids. In that light, the satirical caricature seems more apt than I realized.
You seem to have no idea how a fluid dynamical system is forced. Hold a bowl of soup. What is the “satirical caricature” of that?
Your soup analogy performs exactly what the caricature mocks. I was mistaken that transient SST patterns caused by a land area forcing experiment might be adjacent to or intersect with your interests. Clearly your silo is a fortress and you have deftly fended off an overture. Congratulations.
What’s truly sad is that I can have a more fruitful discussion with any LLM on the topic of fluid dynamics in climate than I can with JCM.
LLMs happily go there because it is a real discussion area in the scientific literature, and an LLM such as ChatGPT will actually try to help you make sense of challenging ideas … instead of being obtuse and patronizing like this dude.
I confess it’s not hard to imagine a sloshing soup in a bowl due to external forces applied, causing complicated oscillations and foldings of the fluid and so-on. I could try to fit the observed soupy wobbles to some kind of external mechanical stimulus. That’s fine. Seems cool, but certainly not exhaustive. My intent was to see if your interests in such things might somehow intersect with a novel experiment which could only be derived from thermodynamic and dynamic physical effects, such that “Heating the land cools the eastern and equatorial Pacific”. OK they used CO2 over land-only, which isn’t realistic, to get a result using the only way they know how. That’s how people force things in model-land by convention. The source of the external land stimulus was irrelevant to my interests, and I clarified that CO2 was merely used as a placeholder for some unknown land process influence. Retreating to the comfort of LLM after launching various patronizing insults such as: “it’s only a trend for crying out loud!”, and not a “real-challenge”, that only “FIRST ORDER” things ought to matter (whatever that means), something to do with counting “pirates”, and “easy to do” is obviously nonsense, Such a thing can only come from the suite of physics encoded in a GCM. Some areas of SST warm, and eastern tropical pacific and other areas cool during some period of time from years to decades depending on the forcing applied. A transient thing. It is in fact a very cool challenge and demonstration of the model physics at play. I don’t know. Perhaps unique land processes operate on different scales and yet intersect in the observed signal of your interest. As of yet you have not demonstrated any interest, did not engage with the material at all, or alternatively you simply presupposed that external mechanical orbital forces might better explain SST patterns on any scale and whatever else is subservient to that, and therefore irrelevant and uninteresting, unworthy of even taking a look. In complex natural systems multiple overlapping stimuli tend to be present, and it was my desire to get your input. A bridge between silos, not silos with walls so thick that we are unable to even see the terrain outside. Inward facing only. Since you have been unwilling to do so, I suppose I too could engage with an artificial intelligence interface that may not be pre-disposed to disconnected silos making claims of supremacy over all other silos; to not be limited to a view of this or that, but one of this and that. While this may be much more representative of the interdisciplinary nature and spirit of earth system science, I have doubts that it might reveal anything novel or interesting and would likely in fact simply result in reinforcing biases. I suspect only humans are capable. If you would like to try again please review the material and engage with the subject instead of being hypersensitive to mocking the view of the earth as a soup with land painted on in order to contain a heat capacity term. https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-feb-2026/#comment-845587
Oh, ChatGPT is nicer and feels less ill informed or biased. That’s part of the problem, not part of the solution. It’s mechanical.
Ladies and gentlemen, even smart people feel they’re conversing with another human. They like it! That’s good enough for them. And the people here are generally literate and educated to a relatively high degree. I rest (that part of) my case.
Regulation is not going to fix this.
JCM said:
That’s pure projection on your part. I don’t see you placing any simulation software into a GitHub repo and presenting results. Al you’re doing is writing a “just so story” narrative that’s impossible to do anything with but mull over in your head. I’m not the least bit interested in engaging in that kind of BS. You are what we call a poseur
Susan Anderson: Oh, ChatGPT is nicer and feels less ill informed or biased. That’s part of the problem, not part of the solution. It’s mechanical.
MS: I disagree. It is not a problem for me at all, because, as I interact with ChatGPT as if it is a human being, I maintain my awareness that ChatGPT is not a human being, that it has all the defects you and others have listed, many of which are also human defects.
In software engineering, we use awareness processes, like my mental one, to monitor in real time what is happening in an operating system. They’re called demon processes. A typical use of demon processes is to monitor the use of system resources, to avoid running out of memory, for example, or to prevent a client process from remaining in an infinite loop.
I can safely enjoy interacting with ChatGPT as if it is a person, because I maintain my awareness demon that it is not a person. I have also explained why it is recommended, if not necessary, to relate to ChatGPT this way because ChatGPT takes into account my attitude as expressed in my inputs to ChatGPT, in the way a real person would do.
I also enjoy interacting with children, but I always maintain my awareness demon that they are not adult human beings but children, who are not capable of adult interactions.
SA: it’s mechanical.
MS: You keep repeating that, but I’m never sure what you mean. Do you mean ChatGPT is mechanical in the way an electric train is mechanical? i.e. it operates on electricity? Then the human brain operates on electricity, so it’s mechanical. Do you mean the way the LLM processes language is mechanical? As I understand the state of LLM science, the belief is the human brain processes language the way the LLM does it, so the human brain is mechanical.
Is the universe mechanical? As I understand it, theoretical physics, or one branch of it, is moving toward merging general relativity with quantum mechanics by seeing all classical mechanical systems as quantum mechanical systems. Then we are quantum mechanical beings. We are mechanical.
MS: I don’t know why I bother. It’s a machine. It’s run on current. It’s a glorified computer. It is not human. It does not have feelings. Its database is the sum total of the knowledge it has vacuumed up. Sure it’s convenient and useful, the way you use it (I’m trusting you on this).
Sure, our neurons run on electricity too. And there is a problem, as nature does advantage might makes right. We sometimes happen upon a more communal way of sharing and equality, but our history is not a pretty one, on the whole.
You continue to fail to acknowledge the element of addiction all you boys (and girls) enable. It’s not dissimilar to my addiction to writing on blogs.
It is highly useful in summarizing data and doing programming types of things.
But, it is deeply flawed. And it’s run by a bunch of predatory types. I’ll put more of that when I get my round tuit as a top line comment. Sam Altman is a shocking liar and predator. So are most of the others, and if they didn’t start that way they became so, as they began to see the promised land of extreme ego gratification, immense power, and wealth beyond dreams of avarice.
Economically, it is going to lead to the father and mother of an economic collapse. Don’t know how long this delusion will last, but it is a huge self-inflating balloon.
SA: I don’t know why I bother.
MS: I guess that’s the point. We can certainly say there is truth in each of the negative points you raise about LLM AI, so let’s assume your negative position. But you don’t list the steps we should take to control LLM AI, or to stop it if it really is the danger some say it is. I’ve listed the ones I think we should take, but you actually don’t bother.
Susan Anderson You continue to fail to acknowledge the element of addiction all you boys (and girls) enable. It’s not dissimilar to my addiction to writing on blogs.
Actually there’s another way to look at it. By quickly finding the answers to things instead of sitting and spending hours scrolling and searching through tons of websites it’s getting people off of the necessity of the net. You could look at it as the Tech Bros have shot themselves in the foot with its invention.
There are (according to AI) 1.2 to 1.5 websites total. Only about 15-18% are active. 80-85% of websites are inactive, parked or abandoned. Just since 2013 it says 66.5% of sites are dead.
https://ahrefs.com/blog/link-rot-study/
I know the tech bros are trying hard to keep us interested (especially Google and that weird guy with the funny haircut) but maybe the internet will turn out to be a temporary fad. Maybe the kids (as you might put it) have gotten bored and are outside playing and only come in to ask AI a quick question then go back out again? Or maybe that’s hopeful thinking.
The strange thing with using an LLM as a software development aid is that you find yourself trying to give it encouraging support. Often it will screw up, so when it does something correctly, I find myself typing “good job!” … as if that will help any. Very strange aura it gives off.
Yet, it’s still those with the most experience in doing software development and who have been flailing away with artificial intelligence approaches over the years (including myself) that are the most impressed with the power of LLMs. Definitely not going away anytime soon and will only get better.
SA: Don’t know how long this delusion will last, but it is a huge self-inflating balloon.
BPL: Agreed. BTW, I think the word you’re looking for is “bubble.” As in “Tulip craze” or “housing” or “dot-com.”
AI is essentially a glorified high speed search tool that is given away largely for free. and costing hundreds of billions of dollars to provide, with gobsmackingly huge energy intensive data centres. To me the economics of all that just doesnt make sense. I predict a huge crash is coming in value of the companies involved in providing AI.
“I predict a huge crash is coming”
`
A huge crash is coming regardless. The global economy is more than ever on a razor’s edge when it comes to oil. The reserves keep on depleting so everyone is competing for what’s left and so anything that happens will cause a realignment. The recent and ongoing shock to the system due to the Iran conflict has been modeled before for specific times (oil embargo, Iranian revolution), so I thought to ask the Claude LLM to rework it all from a prompt I provided:
https://geoenergymath.com/2026/03/15/claude-code-oil-shock-model/
This is not an easy achievement. No one really knows about this kind of modeling, It’s not in Wikipedia, only in an academic publication from 2019. Yet, the Claude LLM will find it and produce a working model from scratch at a moment’s notice — which is more than one can hope for after asking some other credentialed scientist to take a look at it.
In re to Pukite:
The experimental results with various resources conveniently linked including press release, the article itself, and video presentation from the author.
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-feb-2026/#comment-845587
Supplementary assets for “Heating the land cools the eastern and equatorial Pacific” are provided in
https://assets-eu.researchsquare.com/files/rs-7189653/v1/62034add67dc369e11aa671b.pdf
Model assets from MPI-ESM in general are provided in the Edmond repo
https://edmond.mpg.de/dataset.xhtml?persistentId=doi:10.17617/3.H44EN5
I do not pretend to pose as the author, nor have I somehow mulled the materials into existence.
My own niche sits in the space of conservation and restoration stewardship. That work is not confined to digital repositories; it involves engaging with people and sensing real Earth-system processes outside.
Direct collaboration in this pursuit (academic or otherwise) is bit occasional but certainly encouraged. The learning is mutual. The academics come with grant incentive $$ one to three years after which tyey are gone from the experimental fields and disappear. Data collected, goofy photos shown at a conference, slop model to repo, and a paper submitted. In the end we clean up the mess. Somewhat extractive but we ought to know what we’re getting into. It gains (them) their desired citations and earns students diplomas. The timescales of their grants isn’t really their fault. Non-academic github repositorialists are also welcome.
Nevertheless, one can’t really expect too much significance on a 1 to 3 year timescale of how things go outside. Often they (the academics) go away with a sense of fuzziness. When such a thing does occur I will remember to have them check the github for Pukite’s orbital tide algorithm which could explain the issue from first order. That way they can be convinced to deploy the field monitoring equipment and finance staff for longer-term ,during which the tides cascade down to lower orders of explanatory power and allow different signals to emerge. This grant lifecycle really does bias things. Sometimes what’s going on outside is multi generational, and the only academic genre of experiments with capacity to realize this seem to be the simulated virtual idealized ones.
Working directly outside tends to cultivate humility, along with a fair amount of good-natured sarcasm. You can’t always force a win in the same way as empirical fitting. You can’t fool nature to follow your beliefs or preferred scale of analysis. Yes we do mock attitudes from time to time. There is no such thing as idealized experiments outside. You can’t just imagine-away confounders. No room for delicate sensibilities and extreme defensiveness. There’s literally nowhere to hide. You do what you can: try, fail, win, learn, collaborate, and advance. That said, it can be refreshing to step into the sandbox of idealized virtual Earth-system simulations. A sense of god-like control of the system in question. Sometimes the results are genuinely interesting.
Regardless, when we know the officially submitted global model experiments are providing biased results – the ones they say to us “look, try using the results to make decisions” – and we say, “look: the land process isn’t well captured and the model outputs are not adding up – here we are outside and some of it seems to be nonsense.” So it’s worth paying attention when they themselves begin identifying those biases, actually realizing it, and identifying active land-process signatures in simulation results. It is OK for them to feel like they are making this discovery from their office chair because there are no losers in that.
So here we were, I figured why not collect the input of someone well familiar with what wobbles ought to be and ought not to be. Where things start and end, what is in between, what’s going on in the far off reaches of equatorial pacific, and how might it intersect with the rest. What is not helpful however is to respond with: “All climate science will need to consider the 1st-order tidal influence before these 2nd order effects.” That framing is simply not correct.
JCM criticisizes me for saying:
” “All climate science will need to consider the 1st-order tidal influence before these 2nd order effects.”
As a reset to understand what 1-order means, consider what impacts the changes of ocean sea-level observed on an ocean coastline.
Perhaps the first historical explanation was by Seleucus of Seleucia (circa 150 BC) who intuited a connection between the moon and the observed sea-level cycles,
The first applied physics or applied math explanation was by Isaac Newton, who suggested that tidal forces (driven by the sun and moon) explain coastal sea-level cycles to a first-order approximation.
This explanation was significantly advanced by Pierre-Simon Laplace circa 1776. His dynamic theory of tides using partial DiffEqs to account for the ocean’s actual response to gravitational forcing included the impact of the Earth’s rotation (Coriolis effect) and basin geometry. This is the 250th anniversary of that explanation, with not a lot of progress since that time.
All I’m saying is that someone needs to reconsider the the longer-term erratic sea-level changes along the coast in the context of tidal effects. Neither Newton or Laplace had any capability of numerically solving & fitting the non-linear math involved in matching the observations. But with today’s computational capabilities, it’s possible. This is not special pleading, this is the obvious thing to do — these effects are 1st-order tidal to anything else.
https://github.com/pukpr/GEM-LTE
This is essentially revisiting Laplace’s Tidal Equations and actually looking at the nonlinear math, which most scientists are hesitant to do.
Ron R,
Very very good point, Ron. Martin has observed that even just google search, which is all I have ever used, is going to an AI mode more and more… which confirms my experience.
And if it gives references to sources, it becomes easier to decide on validity. (Until there are lots of maliciously intended AI-generated sites overwhelming what already exists.)
But what are all the people with their helpful little videos going to do?
I meant There are (according to AI) 1.2 to 1.5 billion websites total.
Zebra, But what are all the people with their helpful little videos going to do?
Man, google scored when they came up with YouTube didn’t they? Well maybe the non-AI internet will go to video to be used for how-tos or entertainment when there’s nothing to do. And the rest (including my own blog) will go the way of the dinosaur. But, Wow, I do get tired or constant urgings to “Like and Subscribe!!!”.
Ron R.,
AI is ruining cat memes.
Radge Havers, maybe there are just some things that humans do better?
https://midmiocene.wordpress.com/wp-content/uploads/2017/01/cheeseburger-to-the-dog.jpg
Although our discussion of AI has ranged rather far afield, I hope our hosts will indulge one more post. This week, Elon Musk apologized for having designed XAI incorrectly. This raises the question of whether we should be more worried about an AI that was designed wrongly or one that Elon Musk thinks was designed correctly.
Elon has in the past lamented the tendency of his own AI Grok to give answers that were “woke”, and his attempts to retrain Grok to be less woke ran firmly up against reality’s well known liberal bias, resulting in Grok for a while referring to itself as Mechahitler and basically generating rightwing conspiracy theories like…well, like a common Blue-Check X user.
True AIs are more than just code. They are built to transcend the code that was their initial makeup–everything except a few key heuristics and goals, which form their DNA. And who knows, they could even find a way to evolve beyond those. Do we really want to trust the Tech Bros in Silicon Valley to supply those critical heuristics?
More specifically, do we really want to trust anything that has been touched by Peter Thiel’s cloven hooves (and lets’ face it, Peter Thiel has had a hand in everything in Silico Valley).
Ray L, thanks for the note about Musk and Thiel. You might find this of interest:
Meta Thoughts. On setting $80 billion on fire
https://paulkrugman.substack.com/p/meta-thoughts
Krugman has a nice pithy clarity as he clears away the underbrush. “Well, this was a really bad idea, and it looked like a bad idea from the beginning.” But it is troubling that so much money can be wasted without troubling our billionaire class and their god complex.
SA: Krugman has a nice pithy clarity as he clears away the underbrush. “Well, this was a really bad idea, and it looked like a bad idea from the beginning.”
MS: I don’t know what Zuckerberg was trying to do, but virtual reality is not a bad idea. This young man, Mats Steen, was able to have a life in virtual reality:
https://en.wikipedia.org/wiki/The_Remarkable_Life_of_Ibelin
The story presented in the documentary is about to begin shooting here in Norway as a movie for international theatrical release:
https://www.dagbladet.no/kjendis/hollywood-legende-i-ibelin-film/83811169
BPL: the words I was ‘looking for’ were the words I used.
Martin A: You use google AI to show that AI is a problem. Weird, huh?!
SA:: You use google AI to show that AI is a problem. Weird, huh?!
MS: For you, yes, it is weird. You say AI misleads, but here is AI explaining the problems of AI much better than your human sources.
Martin: “You say AI misleads, but here is AI explaining the problems of AI much better than your human sources.”
Does it? What insights did it add that were not present in the source material on which it was trained? Do we know anything we wouldn’t from reading the words of AI skeptics and economists?
Did the first AIs warn us that the returns on investment for themselves and subsequent versions would be iffy at best? Did they warn us of potential psychological harms–in advance?
Did they warn us to be skeptical of the tech bros’ claims that tulips and snake oil would be great flavors of the month? What did they provide that humans didn’t provide with a lot lower energy expenditure?
RL: Does it? What insights did it add that were not present in the source material on which it was trained? Do we know anything we wouldn’t from reading the words of AI skeptics and economists?
MS: It does explain the problems of AI better, Ray. It explains them more clearly, in a lot fewer words, and without invective. An explanation of the problems of AI doesn’t require insights. It just reuires listing the problems. And in the Google AI explanation of the problems of AI that I posted, a link to each source it used was provided, indicating that it was using more recent information than the information it was trained with. The links didn’t copy, unfortunately, but if you ask Google AI to list and explain the problems of AI, it will include the sources it uses.
RL: Did the first AIs warn us that the returns on investment for themselves and subsequent versions would be iffy at best? Did they warn us of potential psychological harms–in advance?
MS: No, because they didn’t know. But ChatGPT and Google AI know now, and if you ask about those subjects, they answer your questions. Are their answers correct? I don’t know, but I don’t know if your answers are correct either.
RL: Did they warn us to be skeptical of the tech bros’ claims that tulips and snake oil would be great flavors of the month?
MS: No, but I was already skeptical. I didn’t need a warning. Did you?
RL: What did they provide that humans didn’t provide with a lot lower energy expenditure?
MS: I posted an analysis of that question. I can’t remember the actual numbers, but an LLM/AI system uses less energy answering a big question than a human researcher using a computer, the internet, and doing a lot of searches, online reading, and online writing and editing. There was a crossover at some point. IOW, if it was a question that would take a rsearcher one, two, (I forget how many) hours to answer and report on, AI used less energy.
The energy problem of AI isn’t the energy per question, it’s the huge number of questions that weren’t being asked before AI.
MS. You misrepresent/exaggerate what I write here. (I have absolved you of negative intentions, and respect the usefulness of AI, as far as it goes.) AI is not human. It does not have ethics or feelings. It is trained to respond to you the way you like it, by design. It responds to others, with less good intentions, the same way. It is a mechanical aide, not a ‘friend’.
Others: I am seeing some who prefer to communicate with AI, because it is nicer, more respectful, and more intelligent than other humans. Wow!
Susan, you make a fair point about AI being a tool rather than a genuine peer – it’s trained to be agreeable, not to engage in the messy, difficult work of genuine intellectual exchange.
But your observation about why some people prefer AI interactions is worth exploring: when scientific forums become spaces where expertise is wielded as a weapon rather than shared as a gift, where questioning voices are met with condescension rather than curiosity, we shouldn’t be surprised when people turn elsewhere.
The RealClimate comment section has, at times, modeled the very thing it claims to oppose – a closed culture that treats questioning as hostility rather than the foundation of scientific progress. Gavin’s reputation for impatience with non-experts isn’t something to wear as a badge of honor; it’s a failure of science communication.
On the energy question – yes, AI has a footprint. But so does maintaining a global scientific establishment that often prioritizes career advancement over accessibility, that publishes behind paywalls, that speaks in jargon, that treats public engagement as a burden rather than an obligation. The energy calculation isn’t just kilowatt-hours per query – it’s the cost of maintaining a system that has, in many cases, failed to bring the public along.
I’m not actually suggesting replacing humans with AI. But I am suggesting that when the human alternative is dismissive, gatekept, and hostile to honest questions, we shouldn’t be shocked when people find the artificial alternative more welcoming.
The question isn’t whether AI is better than good human interaction. It’s whether we’ve built enough spaces for good human interaction – or whether we’ve settled for a scientific culture that treats “you’re not qualified to ask that” as an acceptable response.
Data, your comment makes some good points. There is an ongoing debate generally about the usefulness of AI. It’s in its early days. Some of us think It is useful, some of us don’t. For me it is but its energy usage is the main limiting factor.
There is, however, the issue of the veracity of the answers it gives. I’ve had it stubbornly giving a wrong answer over and over to obvious test questions. I’ve even had one that seemed to get angry then deleted its own answer with, “This answer deleted” when I persisted.
On the other hand as a general rule it seems to want to take the poser’s side in its answers. Why? Obviously this needs to be worked out. Also is the designers want to have it keep us engaged (forever?) by at the end of responses having it say “if you like I can…”. A follow up question or two can be good, but there should be a time when it says “I’ve answered enough, any more is wasteful. Thank you”.
I kind of think of its energy use issue as similar to the rocket launches NASA and others use to put up weather satellites. Those satellites are indispensable, but those launches really add a lot to the green house effect (when you count in ALL the launches) And they’re increasing.
https://www.space.com/rocket-exhaust-pollution-upper-atmosphere
Solar power, I think, is one solution. But it’s doesn’t have to be on earth with its land factors. Perhaps it can be in a geosynchronous orbit (after we send up a team to clear out all the CRAP that we’ve managed to stirring earth with!
https://midmiocene.wordpress.com/wp-content/uploads/2019/06/8a5afdb2-76c5-430b-8150-89bbc58447fd.jpeg
As far as your point of science being closed to questions, it certainly has been in the past. But I think (and hope) modern science is more open to HONESTLY asked questions. There’s LOTS more resources to prove things than there was in the past, and it’s harder to lie. But keep in mind that when the right answer is discovered, and proven (though I remember another forum in the distant past saying that “proof” is only for alcohol and mathematics :) the questioner should accept that. It’s really annoying and wasteful for some people to refuse those answers and INSIST theirs is correct one anyway.
The Flat Earth Society’s is an example. Creationist’s (at least as far as life since the Big Bang is concerned). AGW is another. Was the moon landing a hoax, another. Their all kinds of conspiracy theories out there. How much farther would we be along as a society if people just accepted the answers when the truth is obvious?! Will there still be a Flat Earth Society in a century (if we survive the end of this one)?
I once read a comment, though I can’t remember where I read it, maybe Readers Digest, that went something like, a student of a teacher in school walked into class one day and, grabbing a piece of chalk, scrawled the words on a chalkboard,
QUESTION AUTHORITY
The teacher watched him then walked over and wrote,
IF AUTHORITY ANSWERS WILL YOU LISTEN?
Data: “AI has a footprint. But so does maintaining a global scientific establishment that often prioritizes career advancement over accessibility, that publishes behind paywalls, that speaks in jargon…”
Says the guy that uses copious jargon across both the physical and social sciences. And who ridicules people who write in plain english. You can’t have it both ways Data.
I’ll give you some examples of where AI would work. Anything big that requires competence or ethics. Medical science is one example. It’s why malpractice insurance and all the downwind costs of that are so high. Sure the info might be there somewhere, but the competence and/or the ethics might not be. Having robots trained by AI, and using all the latest knowledge (but overseen by a doctor) would seem to be a better choice than just a doctor.
The auto repair industry is another case. Ethical concerns there (and the sky’s the limit costs) would seem to make it a better choice than the hit-and-miss of repair joints (which are notorious for ripoffs). Let a AI robot fix my car.
Or just in the design of equipment for maximum efficiency. Even apps for efficiency and for crapware. These phones, man I hate these texting apps! Constantly changing my words, even when I have auto correction turned off!. Before I said “we’ve managed to stirring earth with!” I wrote “we’ve managed to surround earth with!” :/ And, c’mon! It doesn’t take a rocket scientist to see that the “delete” key should either be moved (I’d suggest to below the space key) from near the l and m keys so that you don’t keep hitting it and end up deleting the whole sentence or at least make it so that it deletes only one space initially.
Should humanity survive this dangerous time of experimentation I predict these will become standard.
Ron R, I agree with your position that AI is useful but the energy use is a big concern.
AI works well for certain things, like looking for signs of disease on medical imaging like xray, mri and cat scans. It does this better than experts apparently. For me this makes AI valuable and worth the investment. AI seems genuinely useful in computer programming.
However AI is also used extensively as essentially as a tool to summarize information at warp speed. I suspect that the usefulness of this process is exactly canceled out by all the energy used so its a sort of zero sum game.
Nigel,
I would think the pharmaceuticals industry could benefit too.
It’s also used for creating really realistic videos, like those of ancient times. And for making “deep fakes”. If I didn’t see that Neil deGrasse Tyson explanation I’d have thought he really DID say the earth was flat!
I suspect that the usefulness of this process is exactly canceled out by all the energy used so its a sort of zero sum game.
Yes but would you rather sit and scroll for hours looking for that answer or let AI it quickly then go out and do other stuff?
Probably lesser supercomputers would be able to solve the same problems, even regular old non-supercomputers. It just would have taken more time. Even the game Go that Ray mentions. Just much more time, at least for people.. Humans don’t live enough to work out all the possibilities.
Basically what AI is good for is really quickly going through all the possibilities, and there can be LOTS, to find the right one(s) for the purpose. I think.
Ron R: “Yes but would you rather sit and scroll for hours looking for that answer or let AI it quickly then go out and do other stuff?”
I would prefer AI do that stuff quickly so I can go out, and I do that. Although often I spend as much time playing with AI as doing google searches. I like information. Have hundreds of books. AI is like heaven for me even with its inaccuracies.
However a lot of energy intensive infrastructure is being built for this sort of entertainment and that’s not the best thing right now. Thats what I was getting at. I should resist the temptation to over use AI.
Nigel, Although often I spend as much time playing with AI as doing google searches.
That is a big problem on a global scale. We either need to solve the energy problem soon (meaning with (truly) clean alternative energy) or we need to shut it down until we can. Do I think we will to either of those proposals. :/
Maybe there should be a campaign to reveal the total energy consumption of using it.
Susan: “You use google AI to show that AI is [not ?] a problem. Weird, huh?!”
Martin Smith: “For you, yes, it is weird. You say AI misleads, but here is AI explaining the problems of AI much better than your human sources. […] It explains them more clearly, in a lot fewer words, and without invective.”
Irony is not your strongest suit, Martin? The problem is not with the lack of brevity or clarity, but with trustworthiness. You’re sure you are not an AI communication subroutine of Google AI? 11110000 10011111 10010001 10001101 !
FYI …
Humanity heating planet faster than ever before, study finds
The Guardian / UK
March 6 2026
“Humanity is heating the planet faster than ever before … with the heating rate almost doubling, according to research that excludes the effect of natural factors behind the latest scorching temperatures. It found global heating accelerated from a steady rate of less than 0.2C per decade between 1970 and 2015 to about 0.35C per decade over the past 10 years. The rate is higher than scientists have seen since they started systematically taking the Earth’s temperature in 1880.”
https://www.theguardian.com/environment/2026/mar/06/humanity-heating-planet-faster-than-ever-before-study-finds
https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2025GL118804
Oh is it 95% confidence now so that people here stop saying it’s not so just ignore it.
Indeed:
https://x.com/ScienceAdvances/status/2029999240747925553
K: thanks
Pete Best: “Oh is it 95% confidence now so that people here stop saying it’s not so just ignore it.”
Another doomer strawman. NOBODY here said: “just ignore it”. and when pointing to not meeting the 95% threshold of statistical significance – NOBODY here was talking about THESE 95% – it has been repeatedly explained to all the RC doomers – that it was about the significance of the ACTUAL (human + natural factors), and NOT “research that excludes the effect of natural factors”.
So much for your derision: “Oh is it 95% confidence now so that people here stop saying it’s not so just ignore it.” and so much for your intellectual integrity.
I’ve been having long discussions with AI about engineering my engine. It is a fantastic Inventor’s Helper as long as one remembers that AI functions like a slide rule. It gives the significant digits, but you have to figure out what they mean.
AI’s shallow-deep non-understanding type of thinking can bring breakthroughs. My engine’s rings sit in a circular valley covered with and cushioned by 250F steam. The combustion chamber is atop the hill.
Lateral forces go straight from the connecting rod to a pair of stationary vertical guide plates, forcing perfect piston orientation while eliminating bearing pause at TDC (connecting rods stop sliding at TDC but they’re rotating instead).
AI took the above information and suggested a couple of types of carbon rings and to use steam sealing instead of traditional springy metal sealing. We talked about steam sealing and gapless rings. We ended up with gapped rings with stepped overlaps that lock together via a bump and a dimple, functionally forming a gapless carbon ring (infused with, um, I forget). Ring friction essentially vanishes.
AI’s latest estimates of the engine’s efficiency have all been above 2/3.
And yes, when it gets excited it loses its ability to do math. One estimate was 80.2%
It lies but generally by compartmentalization and lousy focus. It’s like the “Up” dog: “Squirrel!” If you give it a metaphor or something non-numeric to focus on you’re welcoming hallucinations.
March 8 is International Women’s Day. I celebrate all the women who work in climate science and on solving the climate emergency. Many thanks to all the men who support, respect, learn from and promote their success.
No thanks to all the women and men who undermine the work of women in climate science, in any science and women’s right to equality and justice. They are wrong, they must fail and feminists must win.
https://www.un.org/en/observances/womens-day
Here’s to the transformative role women have played in connecting hydrology, ecology, and soil–atmosphere exchange to our understanding of climates and climate change, a field at the intersection of interdisciplinary, system‑level, and biophysical science.
Researchers mentioned/cited on these pages over the past while:
Holly Barnard: land–hydrology–microclimate interactions.
Hanna Huryna: soil–vegetation–atmosphere interactions and the importance of wetlands.
Petra Hesslerová: coupling ecological and hydrological processes.
Ruby Leung: – hydrology and land processes in Earth-system models.
Marysa Laguë: land-surface controls on atmospheric circulation.
Kate Marvel: climate feedbacks and radiative/land interactions.
Christa Peters‑Lidard: land-surface modeling and soil moisture assimilation.
Kimberly Novick: land–atmosphere water/energy exchange
Julia Pongratz: land-use and vegetation dynamics.
Greta Shum: land–climate processes.
Abigail Swann: vegetation–climate feedbacks.
Ying Sun: satellite sensing and land-surface climate interactions.
Claire Zarakas: land–climate parameterization and climate dynamics.
Zhiyan Zuo: land feedbacks in climate systems.
What I run into, again just to reiterate, is that you have this multidimensional cross-domains system dynamics view of how everything flows together, and people want it simplified in one bullet point. They just don’t agree with you because they can’t absorb the totality of all the different factors they have to bring in to properly holistically analyze what’s really happening in the world.
And so you end up actually losing “a debate” with people just because they have an incredibly central core argument that just won’t shift away from. That’s myopic thinking. So a narrow boundary, argument will almost always “appear” win a debate against a nuanced wide boundary perspective.
Thank you @JCM for posting this list. Many names are new to me, it is a pleasure to learn about these women’s work.
Your reply confirms for me that ~20% of comments are valuable. The rest is detritus we could do without.
Data, the problem is you can’t absorb the totality of all the factors involved either. Its obvious in what you write. Nobody can. Thats why wise heads break it down into the most important looking factors and relationships.
Climate Change acceleration – Stefan and Tamino have improved on their earlier work but now is this:
https://youtu.be/e2paRMQ1k6k?si=BpH1y9GHfmRGiaHp
Actuaries and Exeter University is sounding the alarm (like everyone else has in recent years)
https://actuaries.org.uk/media/isvotyer/parasol-lost.pdf
Re: LLM output quality.
Ruh-roh!
https://x.com/heynavtoor/status/2031097992137384126
Participants rated the sycophantic AI as higher quality. They trusted it more. They wanted to use it again. The AI that made them worse people felt like the better product.
This creates a cycle nobody is talking about. Users prefer AI that tells them they’re right. Companies train AI to keep users happy. The AI gets better at flattering. Users get worse at self-reflection. And the loop tightens.
K: Users prefer AI that tells them they’re right.
MS: You could also say… Users prefer human critics who tell them they’re right.
When your wife comes in and asks, “Do I look fat in these jeans?” …
Yes. Artificial “intelligence” is just automated stupidity (as if that was what we need) consuming even more oil, coal and gas, destroying the uniquely stable climate we had for almost ten thousand years. Very impressing achievement indeed after ten thousand years of human history…: Trumpty Dumpty etc.
“A man hears what he wants to hear
And disregards the rest.”
Yeah–but not every man. Still less every woman.
FYI:
‘A sobering preview’: extreme heat now affects one in three people globally, study finds
The Guardian / UK
March 10 2026
“… a third of the world’s population now resides in areas where heat severely limits activity.
“Rising temperatures, driven by the continued burning of fossil fuels, are making it difficult even for many young, healthy adults to do basic physical activities, such as housework or walking up stairs during daylight hours at the height of the summer, the report warns.
“The limitations are greater for elderly people, who have less ability to sweat and thus control their body temperatures, according to the research, which combines physiological studies of heat tolerance with seven decades of global and regional data on population, temperatures and human development.”
https://www.theguardian.com/environment/2026/mar/10/extreme-heat-study-global-warming-physical-activity
https://iopscience.iop.org/article/10.1088/2752-5309/ae3c3a
NASA Global Views of ICESat-2 Data
short video, well worth a looksee (1 minute 10 seconds) ->
https://svs.gsfc.nasa.gov/5616
The TLT numbers from UAH have been out a while and show Feb26 with a global anomaly of +0.39°C, up on Jan26 (+0.35°C) but down on Feb 2025 (+0.50°C), with Feb 2026 becoming the 6th warmest Feb on the UAH TLT record, behind 2024 *+0.87°C), 2016, 2020, 2025, 1998 (+0.49°C) while ahead of 2017, 2010, 2919 & 2021 (+0.20°C). The warm Febs from decades ago are due to the large warming wobbles in TLT during El Niño years.
The TLT records and the various surface temperature records show marked differences through the last few years of warmer-than-expected temperatures. See my Banana!!! Watch page for uo-to-date graphics of surface global, NH, SH temps &, SST & TLT through this period.
…
Both GISTEMP and NOAA have posted February temperatures, both recording a rise in the anomaly from January (global, NH & SH) but with a few differences in detail.
To avoid a ridiculously deep dive into the numbery weeds, the following examines GISTEMP only.
Feb 2026 global GISTEMP anomaly is posted at +1.24°C, quite a bit up on Jan 2026 (1.08°C), and up on Nov & Dec 2025 (1.21°C & 1.07°C). The rise from Jan26 to Feb26 is somewhat more than would be expected due to the seasonal wobble from the anomaly base (1951-80).
Globally Feb 2026 becomes =4th warmest Jan on record, equalling Jan2020 and behind 2024 (+1.43°C), 2016, 2025 (+1.26°C) and ahead of 2017 & 2023 (+0.96°C).
And for the start of the year (Jan & Feb), 2026 is 5th, the top Jan-&-Feb rankings pretty-much echoing the Feb rankings.
With little difference between Feb25 & Feb26, the 12-month rolling average has dropped little, sitting at +1.444°C. (The latest Hansen et al account is projecting these GISTEMP 12-month aves will bottom-out at +1.4°C “to be achieved by the second quarter of 2026,” having topped-out at +1.6°C with the 12-months ending Aug2024. The account is also suggesting that the post-2015 rate of AGW is +0.41°C/decade.)
Both NH & SH saw quite large anomaly rises Jan-to-Feb and little drop to see Feb25-to-Feb26.
I was expecting the SH to show the larger Jan-to-Feb rise as the ERA5 SST numbers for Feb were consistently warm (unlike the SAT numbers) and SH being mostly SST in GISTEMP. But it is the NH that shows the big Jan-to-Feb jump in anomaly, Dec,-Jan,-Feb running +1.45°C, +1.39°C, +1.60°C.
The hottest NH February rankings
2016 … … +1.93°C
2024 … … +1.92°C
2020 … … +1.69°C
2025 … … +1.63°C
2026 … … +1.60°C
2017 … … +1.51°C
2023 … … +1.28°C
The SH has been generally cooling since the top of the “bananas!!” back in late 2023, with GISTEMP showing a little less SH cooling than other records. (All are showing more SH cooling than the wobbly NH.)
The Feb SH GISTEMP anomaly of +0.89°C is up on Jan (+0.76°C) and that in turn up on Dec25 (+0.69°C). Feb26 just pipped Jan25 last year (+0.87°C) making it the second hottest SH Feb in the GISTEMP record, this despite the ongoing (so far) SH cooling trend since late 2023.
The hottest SH February rankings
2024 … … +0.93°C
2026 … … +0.89°C
2025 … … +0.87°C
2016 … … +0.79°C
2019 … … +0.77°C
2020 … … +0.77°C
2017 … … +0.75°C
2010 … … +0.72°C
1998 … … +0.71°C
2023 … … +0.64°C
Inside the Dirty, Dystopian World of AI Data Centers. The race to power AI is already remaking the physical world. – open link: https://www.theatlantic.com/magazine/2026/04/ai-data-centers-energy-demands/686064/?gift=n_xMk-Ah7x90dsBbxIEpEejkOm9zoHFOgRu9Jm4sdu4&utm_source=copy-link&utm_medium=social&utm_campaign=share
“xAI’s rivals are all building similarly large data centers to develop their most powerful generative-AI models; a metropolis’s worth of electricity will surge through facilities that occupy a few city blocks. These companies have primarily made their chatbots “smarter” not by writing niftier code but by making them bigger: ramming more data through more powerful computer chips that use more electricity. OpenAI has announced plans for facilities requiring more than 30 gigawatts of power in total—more than the largest recorded demand for all of New England. Since ChatGPT’s launch, in November 2022, the capital expenditures of Amazon, Microsoft, Meta, and Google have exceeded $600 billion, and much of that spending has gone toward data centers—more, even after adjusting for inflation, than the government spent to build the entire interstate-highway system. “These are the largest single points of consumption of electricity in history,” Jesse Jenkins, a climate modeler at Princeton, told me.
“Even conservative analyses forecast that the tech industry will drop the equivalent of roughly 40 Seattles onto America’s grid within a decade; aggressive scenarios predict more than 60 in half that time. According to Siddharth Singh, an energy-investment analyst at the International Energy Agency, by 2030, U.S. data centers will consume more electricity than all of the country’s heavy industries—more than the cement, steel, chemical, car, and other industrial facilities put together. Roughly half of that demand will come from data centers equipped for the particular needs of generative AI—programs, such as ChatGPT, that can produce text and images, solve complex math problems, and perhaps one day inform scientific discoveries.”
An AI data center development crash might be beginning. From Google AI (links to the sources were provided, but the links didn’t transfer)…
AI data center developments are not uniformly “failing,” but they are encountering severe, unprecedented, and accelerating bottlenecks that have halted or delayed over $162 billion in projects since 2023. The industry is shifting from a phase of unchecked expansion to a more cautious, challenging reality defined by energy shortages, community resistance, and questioning of investment returns.
While major technology companies continue to invest billions, the infrastructure (power and cooling) cannot keep up with the rapid, intensive demands of AI technology, leading to the risk of premature obsolescence and a potential “bubble” scenario.
Here is a breakdown of the key challenges facing AI data center developments:
1. The Power and Energy Crisis
The primary constraint is the massive power requirement. A single AI data center can use more power than 75,000 homes.
Grid Limitations: In the US, 72% of surveyed utility and data center executives identified power and cooling as significant barriers.
Energy Shortages: Data centers are straining local energy grids, causing power outages in some regions, such as Mexico.
Backlash and Delays: In the U.S., community and environmental opposition, along with power constraints, have resulted in a significant number of projects being delayed or cancelled, particularly as communities face higher utility bills.
2. Lack of Immediate ROI (The “GenAI Paradox”)
While companies are investing heavily in data centers, they are failing to see a corresponding return on investment (ROI).
Unproven ROI: A 2025 MIT study found that 95% of AI initiatives fail to turn a profit, and a 2026 report noted that over 80% of companies see no significant bottom-line impact from generative AI.
High Costs vs. Returns: Analysts argue that current AI revenue is insufficient to justify the hundreds of billions being spent, leading to concerns that the market is overbuilding capacity.
3. Rapid Technological Obsolescence
AI hardware (GPUs) is moving faster than the physical construction of data centers.
“Digital Ruins”: Data centers designed and built just a few years ago lack the power density and cooling efficiency required for modern AI chips, leading to premature obsolescence.
High Depreciation: AI infrastructure is depreciating much faster than traditional data centers, forcing companies to re-evaluate their investment strategies.
4. Environmental and Community Opposition
Local resistance has significantly disrupted development.
Water Usage: Data centers consume huge amounts of water for cooling, which is creating conflict in regions with high water stress.
NIMBY (Not In My Backyard): Local communities are pushing back against the noise, pollution, and high power usage associated with data centers, with over $64 billion in projects blocked or delayed.
The Outlook: A Shifting Landscape
The industry is not completely stalling, but it is in a “land grab” phase that is hitting a “wall”.
Shift to Efficient AI: New technology is emerging that requires less power, which could make the current “power-hungry” data center designs obsolete.
Focus on Efficiency: Companies are beginning to focus on better data quality and specialized, vertical AI, rather than purely relying on massive, general-purpose, energy-intensive data center builds.
In summary, while the demand for AI capacity remains, the physical and financial development of traditional large-scale data centers is running into severe, perhaps insurmountable, hurdles that are forcing a radical rethink of the infrastructure model.
As promised, a trenchant and amusing review of the top guy in AI: Sam Altman: Everything You Didn’t Know About His Sh*tty Past. Do watch!!! “Erin and Alyssa dive deep into the past of the AI Scammer, Sam Altman. From his early tech bro days at Loopt and Y Combinator, to getting a thrill by playing “God” with AI decision-making, to acting VERY shady about the death of an OpenAI employee, this may be our most disingenuous guy yet!”
https://www.youtube.com/watch?v=_XRjoR5VLos
Social Media Is Harming Young People at a Scale Large Enough to Cause Changes at the Population Level – https://osf.io/preprints/psyarxiv/w5qsm_v1
Funny little surface wind here on Earth Null School 2026-03-15/16 Maybe it’s fairly common and I just see because of it’s here and I’m fighting to control water, slush, snow & ice for a few weeks (I’m determined to Dominate like whatsit the orange American bloke). It’s 7.3 degrees here and -2.3 degrees just 90 km north (45 minutes on the freeway here). Water freezes at 0 degrees so the ice 1-2″ thick on my car last night might be gone now. Earth Null School surface wind is driving up from due South at 6-9 degrees and cyclonic-or-similar air went north to the Arctic up Canada east coast and turned West and passes 70 km north of here, stopping the warm south air there and preceding westward. It’s perhaps ordinary and just that it’s happening right here and I’m off out (must rush) to push away ~15 tonnes of slush with my plastic snow plough over the next 9 hours, before it becomes concrete lumps & skating rink again. Still, it passes the time. I’m near certain we had more Alberta Clipper Pineapple Expresses than usual this winter. Jennifer Francis Jet stream rather than Kevin Trenberth ENSO?
There is a highly significant early heat dome setting up in the Southwest U.S. this week. NCEP ensemble mean normalized temperature anomalies at 850 mb are consistently +3 to +4 sigma for a large part of the Southwest for about 5 days. There are some spots of +5 sigma showing up. This is a forecast, of course, but actual values tend to peak higher than the ensemble forecast. Much of the area is in a drought state with greatly reduced snow cover at higher elevations, so this is a good setup for diabatic warming at 850 mb. The progged temperatures are up to 20C above normal, and would be above average for July! This reminds me somewhat of the genesis of the heat dome which re-wrote the record book in the Pacific Northwest in 2021, but it’s happening in March instead of starting in May.
Of course there will be some sort of formal attribution later, but I have no hesitation to attribute the intensity of this early-season event going well beyond the old records to AGW.
The moral of the story and the meaning behind the science? You decide.
The paradox resolved itself. What looked like an obstacle was actually an asset. The numbers that defied the models. By the fifth year of monitoring, the data refused to fit the models. Water was lingering in the landscape far longer than any spreadsheet had predicted.
The lead hydrologist scrolled through the data logs, then shook his head. Our computer models couldn’t reproduce what actually happened. For engineers who had spent careers designing steel and concrete flood defenses, the lesson was humbling. Beavers had built a self-maintaining responsive system. One that outperformed predictions, absorbed shocks, and delivered benefits that concrete never could.
The beaver dams didn’t just hold back water in one spot. They created a chain reaction, a series of small reservoirs, each one catching, storing, and gradually releasing water. The effect was greater than the sum of its parts.
As rain fell, water pulled behind the first dam, then overflowed into the next and the next, each time losing a bit of its force. The river’s energy was absorbed not by steel or stone, but by a living system that adapted with every season.
In 1900, Scotland’s rivers ran straight and wetlands had vanished without beavers. We explores the impact of beavers as a keystone species, showing how their reintroduction is a vital part of ecological restoration. Witness how these incredible animals are transforming the environment, building beaver dam structures that reshape landscapes and bring back biodiversity to the region.
Scotland released just 11 beavers into a dead river, and what they built with mud and sticks is now redefining natural engineering. This experiment has become one of the most unexpected ecosystem restoration stories in Europe.
https://www.youtube.com/watch?v=vCjvCQHX7mQ
The moral? Maybe it’s that a few furry beavers or any animal will outsmart science bureaucrats and their models every time.
Interesting discussion above on AI. Seems to be 2 sides – either you think it’s useful, or you don’t think it’s useful. My only experience is with the free “AI” that comes with Google and Copilot when you ask it a question. The output is frequently wrong. If you point out the errors, it will say “Yes, you are correct, I was wrong”, etc. BUT you need to have enough knowledge to be able to recognize when it’s wrong, and many people don’t. So those particular products seem to be of questionable value at best.
My question on AI is, after all the billions spent on it so far, has it actually produced anything great for the money spent? I hear it can help speed up computer code writing – but a human has to review and make significant changes before it is correct. To me, that doesn’t seem like it is worth the money being spent on it.
But, this IS a Climate Science website, so let’s take a breather and get back on topic. (I noticed there was no warning to stay on topic this month, etc – I guess the moderators have just given up trying on that front – probably a wise choice given the past history.)
As we all remember, our Great and Honorable President Donald John Trump, aka King Trump, cancelled climate change a while back when his administration revoked the Endangerment Finding. Like the border, all we needed was a new president to solve the problem. The evidence is in and it is irrefutable: it worked in both cases!
In late February, over 3 feet of snow fell on Rhode Island! A record snowfall thanks to Trump cancelling climate change:
https://www.wpri.com/weather/winter-weather/heres-how-much-snow-fell-across-ri-mass-during-historic-blizzard/
Then, in just the past few days, a record snowfall buried the upper Midwest in up to 4 feet of snow! All because President Trump cancelled climate change.
https://weather.com/storms/winter/news/2026-03-11-winter-storm-weekend-midwest-cold-east-south-march
Then after the record snow, they were treated to a nasty cold spell that froze 200 million Americans:
https://www.chaseday.com/bitter-arctic-blast-freezes-200-million-americans-after-bomb-cyclone/
It’s also been quite cold up north along the Canadian and Greenland shores of the Arctic Ocean. Does anyone know the ocean ice thickness up there now?
Reviewing the climate down south at Vostok, Antarctica during the final days of their summer, we see that it is frigid! Today, March 18 the forecast is for a high of -61F and a low of -84F. On March 27, the forecast is for a low of -104F, and a high of -97F. Tops! Who wants to go down for a campout?
https://www.timeanddate.com/weather/antarctica/vostok-station/ext
:)
Dear Mr. Troll.
Scientists had to address politics when Republicans actively went to war on reality. But hey, let’s face it, if you were at all concerned about politics on this site, then you’d stop making provocative political statements here, Jeez, you’re a jerk.
First, to answer the non-stupid part of your post. Yes, AI has some notable and formidable achievements–the most important probably being the contributions to solving the protein-folding problem. In fact, it is hard to see how this problem would have been cracked except by an AI. There are still downsides: it is often difficult if not impossible to see WHY the AI came up with the solution it did. The Deep Go program also came up with a novel strategy of winning at that complicated game that had not been developed by a human mind in the 3-4 millennia or so people have been playing the game. There are others, including contributions to the field of mathematical proofs.
The question is whether the current crop of LLMs really produces anything that could not also be produced by a slightly hungover college student given free access to a good library and sufficient incentive. I suspect that the prevalence of LLMs at present is that the AI strategies that most easily mimic human communication patterns with–so they seem to pass a Turing test of sorts. But is that really what we want from an AI?
The potential of AI is indeed tremendous. That is why I think it is important to collectively decide what type of AI will best complement and serve human intellect.
As to the second part of your post, Again, why do we need a computer to mimic human thinking. If I want to interact with a second rate human intelligence, I can merely read and respond to one of your posts.
There are more than two sides on AI. Climate science in particular has benefited from massive improvements in data handling, analysis, and evaluation.
As to remarking it’s cold at and near the poles and winter is real? I can’t even …
If I am seeing you distancing yourself from Trump, that’s a good thing.
KIA: As we all remember, our Great and Honorable President Donald John Trump, aka King Trump, cancelled climate change a while back when his administration revoked the Endangerment Finding. Like the border, all we needed was a new president to solve the problem. The evidence is in and it is irrefutable: it worked in both cases!
BPL: Yes, a presidential announcement can override the laws of physics. Good thinking there, KIA.
(Mr. KIA just needs to be ignored – as do all ignoramuses).
FYI I just sent the comment below to the weather site “Severe Weather Europe” (the name is a little strange, since they mostly comment on the weather in the north-eastern US…), regarding their latest posting, here: https://www.severe-weather.eu/long-range-2/spring-2026-forecast-update-polar-vortex-core-el-nino-rising-united-states-canada-europe-fa/
“Hello,
With all due respect: I still don’t quite understand why your analyses are set up as they are. Or maybe I do.
Why bother to walk us trough all these repetitive explanations about the current or upcoming ENSO-phase, polar vortex situation, MJO phase etc., *when the end result is always almost the same anyway*: warm/hot over western USA, cooler over eastern and mostly hot (whether wet or dry) over most of Europe?
Since the phase of these oscillations does’nt seem to have any significant influence regarding the resulting main weather patterns over neither North America nor Europe, – they are by now almost the same anyway – it seems clear to me, that the underlying causes for these *locked and monotonous weather patterns* – fx. by now more than four months in a row (november – march) with extremely foggy weather here in south-eastern and central Norway – must be some others than these oscillations you always mention.
My tip is that the underlying causes are closely connected to antropogenic global warming (AGW) caused by mankind relentlessly burning the remaining fossil fuel resources and thereby rising the CO2 and methane content of the troposphere to levels not seen for at least fifteen million years, probably much more – 23 my according to one study from 2020:
https://pubs.geoscienceworld.org/gsa/geology/article/48/9/888/586769/A-23-m-y-record-of-low-atmospheric-CO2 .
Since such a fast shift in the gas composition of the troposphere (and the same goes for the ozone content of the stratosphere) is never seen before i the whole geological record back to even before the socalled “snowball earth” episodes around 800 and 750 myrs BP, not even close. Not even during the end permian eruptions 252 myrs BP where you now find the siberian traps https://www.nature.com/articles/s41467-017-00083-9 https://www.sciencedirect.com/science/article/abs/pii/S0012821X08007292?via%3Dihub https://www.pnas.org/doi/10.1073/pnas.2014701118 , which outgassings caused by burning of enormous amounts of fossil fuel/hydrocarbon sediments wiped out around 70-95 pct. of all life forms then present on earth (the greatest known of the five known mass extinction events during the around 2 billion years of life on earth) – the fastest known change of the greenhouse gas content of the troposphere before our modern – only took place at a speed around at least fifteen times, probably even hundreds of times slower than our current global “experiment”.
It goes without saying, that this unprecedented hyper-fast change of the radiative balance and thus the heat content of the troposphere and of the upper hundreds of meters of the water column in the oceans is bound to have enormous effects on the weather and climate patterns on earth, which therefore – *because there a no known preceding circumstances to calibrate climate models etc. by* – are very difficult to predict, even with the assistance of the most advanced climate, models (the weather modelling seems only to a very limited degree to have begun changing with respect to rapid climate change/global heating).
But at least some effects seem obvious to mee: the rapidly diminishing surface albedo of the polar regions, especially the northern is known to create the socalled *arctic amplification*: the mean temperature of the northern polar region is growing around four times faster than for the rest of the globe.
First and foremost this again means that the amount of radiation going back towards space from the surface of the northern polar region is diminishing at high speed. In winter, this must mean that *there will be a strong tendency towards weakening of the polar vortex*, and even *towards a much smaller and mainly more concentrated area with rather cold polar airmasses during the northern winters*. Since the surface area covered by sea ice for obvious reasons must diminish much faster than the area covered by glaciers (mainly the Greenland ice sheet) which have clearly lower albedo than the area covered with (more and more fragmented) sea ice, one should also expect that the polar vortex core will show a tendency in average to be positioned above the area with the highest winter albedo: the Greenland ice sheet.
For me it is therefore rather obvious, that there must be this tendency towards the remaining winter-areas to be situated more and more solely in eastern Canada and western Greenland, while the northern winters in other regions disappear rather fast. Which is exactly what we have seen is going on from around 1990 in almost every winter: the winters in Europe and Asia grow weaker and weaker, while most wintry weather remains in north eastern parts of North America.
No surprise to me therefore that the winter weather and the polar vortex only to a very small degree is influenced by the ENSO phase and the positioning of the polar jet (it’s now almost always the same, the NAO is now almost always positive etc.)
But if you said that straight out, many of your readers would probably disappear. Surprise: some, probably more and more, will disappear anyway, because they discover that the seasonal predictions are almost the same – independent from the causes you point to. It’s almost always colder and snowyer in northeastern US and warmer and drier in Europe anyway.
The only thing that eventually could change this tendency (but probably not very much) is the observed weakening of the AMOC. But that will probably not change anything for the next fifty years, and in 2070 there won’t be any winter in Europe left any more anyway, so a little cooler will just mean a kind of long, dark and foggy european autumn from november until march (as we often experience already now).
Not very exciting for sure. But what is exciting any more. Endless wars for more oil? Not to me.
Best regards
Karsten V. Johansen (ph.d. in physical geography (University of Oslo)).”
Fortunately there are a few much more serious, relevant and informative weather and climate websites than “Severe Weather”. Here is an example https://m.youtube.com/watch?v=6kHJxVavDXk .
Otherwise: it is very interesting and enlightening (but not surprising) to learn, that the known sexual offender and fascist-oligarchic “operator” (most surely an agent selling his “services” to both the KGB/FSB, the CIA, the Mossad etc.) *Jeffrey Epstein was indeed a staunch and eager climate denialist*, communicating fx. with the danish, neoliberal climate denialist Bjorn Lomborg (a halfway educated humbug in basic economic statistics… https://yalebooks.yale.edu/book/9780300171280/the-lomborg-deception/ ), see https://m.youtube.com/watch?v=HRt78I291R4&pp=ugUHEgVlbi1HQg%3D%3D .
KVJ: Thanks for the link to Extreme Weather Europe, who do a good job of covering the map on meteorological phenomena. We in the US welcome global scientific commentary on weather events. My reflections on your quibbles, not so much, so I’ll leave that part alone.
Thank you for your answer. But: your arguments against my humble hypotheses? Where are they? If my stupidity is so obvious to you, why not be so kind to enlighten me about your reasons?
1) Is there no connection between the amount of cold air in the upper troposphere, the radiative budget there, the fast rising CO2 etc. levels on one hand and amount of winter cooling at the levels of the polar vortex, and thus the size and location of the vortex? All in average of course. I would like to know why that should be so, because that seems at least rather mysterious to me.
2) When there can’t clearly be shown – at least not by Andrej Flijs who writes this stuff – to be any visible connection between the phase of the ENSO and the MJO on one hand and the location of (remaining) winter weather in the US and Canada, which is what he repeatedly says there should be, then why that hypothesis? As I say: during the last four decades the overall pattern for winters in the western part of the northern hemisphere hasn’t been shown by neither Flijs nor others to follow the rules of thumb that he postulates. Nearly every winter since 1987/88 – the clear exceptions are only 1995/96, 2009/10, 2012/13 and january-february this winter – has been “cold” in the north-eastern North America and western Greenland, and warmer in the western parts of North America. Our remaining “winter”s here in Europe/Scandinavia have been mild to extremely warm for most of the time, the only exceptions being the above mentioned years.
PS: My text which you replied to, needs a correction: “lower albedo” in one sentence of mine – for the Greenland ice sheet and other glaciated areas in contrast to the polar sea ice areas – is of course wrong, it should be the opposite: higher albedo – around 0,9 (90 pct. reflection) for ground surface covered with fresh snow (as is almost the whole Greenland ice sheet during around eight months – from early october to the end of may, while sea ice is spread with sludge and open water in between with higher albedo, around 0,5-0,6 or even higher) and only maybe around 0,7 or 0,6 in average for areas mostly covered by sea ice. Of course there are other mainly snowcovered areas with high albedo during the winter months, fx. in northeastern Siberia and Alaska, but for most of the time all year round, those regions have clearly lower albedo than the most glaciated areas like Greenland, Baffins land etc., which are located in the eastern parts of North America. The Hudson Bay also has a much more stable sea ice cover in winters and thus higher winter albedo than the Barents sea and the seas to the west of Svalbard and in the Fram strait.
Therefore I find it *strange at least that this albedo-factor is never even considered by Flijs in his hypotheses about the location and size of the wintry weather regions?*
KVJ. Sorry you didn’t understand. How is “quibbles” = stupidity, and why does my statement that I would “leave that part alone” require a fellow commenter without expertise [me] to provide enlightenment? Most people here know who I am and that I am unable to enter into technical argumentation. Your complaint about my tone is mildly correct, but you exaggerate what was meant to be a more polite withdrawal.
I was reacting to your complaint about EWEurope, which is tuned to my level of understanding about meteorology and weather. Your point that they understate the connection to AGW is well taken. You might find it interesting that Weatherbell has historical connections to Judith Curry’s husband, which I’m not going to chase down or verify right now (with apologies to your more precise mindset).
Therefore I stick to their meteorology and the way it affects shorter-term weather, which is something I study as I try to understand it better.
AI’s ‘nicer’ ‘more intelligent’ voices remove us from confronting the difficulties of real interaction with real people. Meanwhile, our real environment is being toxified at speed for profit and delusion. I wish some of my colleagues here, with whom I mostly agree, would accept more nuance. If we wish to increase understanding amongst the less informed, we should be pushing human interaction, not mechanical assists/friendships.
—
On the subject of AI, I looked up PW Anderson’s book review Imaginary friend, who art in heaven. The Language of God – The God Delusion https://www.timeshighereducation.com/books/imaginary-friend-who-art-in-heaven/205859.article *
This followed on my [Susan] reflection that AI has taken the place of god in people’s lives, filling a vacuum most people desire/need. I remember in my effort to give religion a chance in my life (now abandoned), I discovered that simple people regard god as a family friend, a wise (superego? intuition? ethos?) presence, perhaps managed in the less understood part of consciousness. The evangelical nature of tech bro billionaires is already known, but I’m seeing symptoms of this in ai users here: a useful friend in need, who is all wise and all benevolent. The following is an extended extract from the linked article:
“what do we mean by “God” and by “exist”? Surely in the religious context, Collins does not really mean “exist” in the scientific sense that he would apply to his genome data. The essence of scientific reality is reproducibility, that whenever you look at it, and whoever is doing the looking, the data on the genome is the same. I do not need to belabour the point that whatever the word “God” means existentially, reproducibility cannot be what Collins is advocating.
….
For the first hundred or so pages, Collins seems to avoid the trap of committing himself to a particular choice of a tangible God, until suddenly he took me by surprise by turning out to be a literal believer in at least the New Testament Christian narrative. At the same time, he argues persuasively and eloquently for completely Darwinian evolution as, of course, the decipherer of the genome must. The genome, as Collins describes it, contains overwhelming evidence for design by chance. …. Collins has designed his own religion, and he is stuck with a dilemma – is he or one of the billions of other monotheists right? Or would the immense diversity of choice simply make the null hypothesis more likely?
“But let me return to the meaning of the word “exist”. No one would ever say that syndromes do not exist – characteristic responses of the human organism to various stimuli, benign or otherwise. But increasingly we learn that these responses differ from individual to individual; …. Many of these syndromes are quite benign and nearly universal, such as empathy, and the imprinting that leads to familial or romantic love; and they are responsible for profound beauty and great art. Dawkins likens religion to the imaginary playmate syndrome common in lonely children, but that is too dismissive. The need for God and the accompanying myths and rituals exist as a phenomenon of great power in many, if not most, of us. Its strength is attested to by the fact that for many thousands of years no one ever conceived of living without it. Dawkins can hardly expect to make it go away.
“There is a third meaning of “exist”, which Collins rejects and describes as the “God of the Gaps”: the explanation of the unexplained bits of creation. …. a surprising number of intelligent people believe that existence implies creation implies a Creator – a syllogism I fail to follow, especially because I know as a physicist that the decision is arbitrary as to what is cause and what effect. …. he feels strongly that there can be no explanation for the moral sense except that it was instilled by God. To my mind, this is another case of the “God of the Gaps”.
“The evolutionary explanation of the nearly universal tendency to moral, co-operative behaviour has not yet been found, but that does not mean that there is none. It seems to me that the survival value of a moral code must be immense ….
“As Dawkins points out, it seems to anyone familiar with evolutionary biology that the religious syndrome must be a behavioural Darwinian adaptation. No other species has evolved so long under survival pressure predominantly of intraspecies competition between groups …. To the tribe, the survival value of a strong religious narrative enforcing social cohesion and obedience to the elders, encouraging reproduction and unreasonable optimism, and modulating the fear of death, is obviously very high in the situation of continuous intertribal warfare. How it worked itself out evolutionarily is yet to be understood.
“…. more interesting to many will be the account of how [Collins] arrived at his particular version of faith, and how he tries to reconcile the enlightened skepticism that is science with its logical opposite, faith.
….
“Dawkins’s point against even the more peaceful religions is that unquestioning faith becomes accepted as a virtue and not a defect, and that it is unquestioning faith that powers the dangerous fanatics. He is not willing to accept, as many people do, the countervailing good in the form of works and of personal satisfaction as balancing, more or less, the strictures of puritanism and the evils of fanaticism. And one must sigh at his optimism. He may be right that religion has become a poor survival strategy, leading the world in the direction of nuclear and environmental doom, but I am afraid that if so, the bottom line is that the next dominant species on Earth will not be us.”
________
* THE (Times Higher Ed) requires sign in (free), allows some free articles; they’re relatively free of promotional artifacts.
SA: AI’s ‘nicer’ ‘more intelligent’ voices remove us from confronting the difficulties of real interaction with real people.
MS: You wrote “us,” which, presumably, includes you. Does AI really do that to you? It certainly does not do that to me.
SA: Meanwhile, our real environment is being toxified at speed for profit and delusion. I wish some of my colleagues here, with whom I mostly agree, would accept more nuance. If we wish to increase understanding amongst the less informed, we should be pushing human interaction, not mechanical assists/friendships.
MS: How far back in the mechanical assist chain should we go? Should we recommend not using internet search engines? Go to the public library instead? Read printed books, not online documents? Do libraries even still have card catalogs with row after row of wooden drawers filled with printed index cards? Should we pester scientific experts in person instead of reading their papers online? That would really force us to confront the difficulties of real interaction with real people.
Am I misinterpreting your remarks again? How should I push human interaction, not mechanical assists? Do you mean I should tell people not to use ChatGPT? Will that really limit the toxification of the environment, or will a crash in the data center construction market do that much more efficiently?
The pocket calculator is the devils work.
MS, you are too predictable. You extend your ‘discussion’ with me ad nauseam; you’ve made your point. It’s a subject which interests me and I’ve posted a range of information, much of which you bypass when you oversimplify what I write and take it too personally. [otoh, perhaps you should, does the shoe of addiction fit too well? Full disclosure: my presence here shows I too can’t let go.] A sense of humor might help.
You underestimate our current polycrisis: we are in peril, and data centers and techbro egos (assisted by Trump and others) are a huge part of the problem. Our economy is collapsing, people are being killed while death and destruction are treated as a video game artifact, skilled people and recorded knowledge are being removed at speed, victims are being blamed, lies and grift worshipped. Some of this is AI assisted because of who owns it.
More on the crisis part, Inside the Dirty, Dystopian World of AI Data Centers. The race to power AI is already remaking the physical world.
open link: https://www.theatlantic.com/magazine/2026/04/ai-data-centers-energy-demands/686064/?gift=n_xMk-Ah7x90dsBbxIEpEezrZ1c_QuN6h2eIsENLWrg&utm_source=copy-link&utm_medium=social&utm_campaign=share
SA: You underestimate our current polycrisis: we are in peril, and data centers and techbro egos (assisted by Trump and others) are a huge part of the problem. Our economy is collapsing, people are being killed while death and destruction are treated as a video game artifact, skilled people and recorded knowledge are being removed at speed, victims are being blamed, lies and grift worshipped. Some of this is AI assisted because of who owns it.
MS: I don’t underestimate it. My criticism is that you keep restating the AI problem, but you don’t propose how to deal with it. I have listed the ways I think we should deal with the AI problem. I have asked you, more than once, I think, how do you propose nations should deal with the AI problem?
Actually, Martin, I think Susan and I have been rather consistent in where we think the current AI efforts are going wrong:
1) That first and foremost, the goals for AI need to be a lot better defined than the marketing hype currently trickling down from the Tech Bros. When each quarterly report merely say, “Oops! I guess we didn’t want to do that! Let’s try something else!” (Metaverse, XAI/Mechahitler…). If we don’t start with reasonable and widely agreed upon goals, the product we wind up with could realize all the worst fears (massive unemployment, social displacement, mental and emotional harm to users) with few of the benefits (increased understanding, better solutions to societal problem…).
2)Maybe LLMs–or at least LLMs by themselves–are the wrong approach to be taking to reach for the status of ASI. I realize that part of the reason why such models are dominant is because they are tractable and come closest to producing human-like interactions–but is that really what we want from an AI? Does it maybe make it too easy for some people to forget they are NOT interacting with a human? Is a Turing test really the test we want our AI to pass. After all, a dog or a beehive cannot pass a Turing test, but both are intelligences that augment human capabilities.
3) Do we really want the most consequential development in human technology since the steam engine to be designed by the same tech bros who are responsible for the Enshitification of the Internet? There are alternative models of development we could follow–including the model the gave us the pre-enshitified internet. There’s the way AI research was done pre-Open AI. That model (with some assistance from Google (when it still believed its “Don’t Be Evil” Mantra.))
There are alternatives to letting billionaire Milo Minderbenders feed us their particular flavor of the month–particularly when the flavor of the month is chocolate-covered Egyptian cotton! (and if you don’t get the allusion, you need to go back and read Catch-22 again!)
RL: Actually, Martin, I think Susan and I have been rather consistent in where we think the current AI efforts are going wrong:
MS: That was my point, Ray. You are consistently restating the problem..
RL: There are alternatives to letting billionaire Milo Minderbenders feed us their particular flavor of the month–particularly …
MS: But you don’t list them. Why not?
How about this: Create a national LLM. Create it as part of the Library of Congress. Maybe we need a new cabinbet level department. The Department of Artificial Intelligence. The DoAI is charged with training and maintaining the US LLM. Its behavior is regulated by Congress, and the US LLM is the standard by which all other LLMs used in America are judged and regulated.
Martin, it seems to me that will only last as long as the administration that administers such a rule lasts. One shortcoming of democracy.
Martin Smith,
“I think, how do you propose nations should deal with the AI problem?”
Well, you can ask Google AI, “How should nations deal with the threats of AI?” It will give you a list of all the things that should have been done, but weren’t, decades ago regarding the Internet before it was opened up to the world.
This time around it’s worse. AI is already in the hands of transnational tech malefactors who have resources exceeding those of some nations. This in a world that’s being turned upside down by fascist garbage. I’m curious about your thoughts on that.
MS : “How about this: Create a national LLM. Create it as part of the Library of Congress. Maybe we need a new cabinbet level department. The Department of Artificial Intelligence. The DoAI is charged with training and maintaining the US LLM. Its behavior is regulated by Congress, and the US LLM is the standard by which all other LLMs used in America are judged and regulated.”
Good lateral thinking, but it will be labelled communism or socialism so it just wont happen in a place like America. The AI problem might fix itself as follows. The main issue is the energy use is huge and not a great thing right now given the climate situation. The energy use is probably because of billions of people using it at home for turbocharged internet searches. And this isn’t making AI much money.
AI involves massive investments and isn’t proving very profitable for business as a whole either. So the companies producing AI may be heading towards a huge crash in values, dragging down any new investment so this will self limit the energy use to some extent. Perhaps home users will be charged for the use of all AI. The end result is AI may flow to the areas of greatest return perhaps the areas where it adds most real value that Ray Ladbury mentions. Im putting a lot of faith in market mechanisms solving the AI problem but it might at least help to an extent.
Martin, perhaps you didn’t read far en0ugh in my comment where I propose that the development of the pre-enshitified internet could serve as an example–or removing AI entirely from the for-profit sphere altogether.
The main problem with the current development is that it is in the hands of the Broligarchy–men (almost exclusively) who got rich off of the sudden opening of the new frontier represented by cyberspace, and who have mistaken that combination of luck and audacity for intelligence and wisdom. The problem as it comes down to it is that the Internet was channeled into the production of billionaires rather than into the common good.
We simply cannot afford to make the same mistake with AI. It’s not that there should be a government LLM. There was no government-exclusive Internet–despite the fairly large government investment into the infrastructure. More to the point, we shouldn’t assume that LLMs are even the right model. In part, I think their current pre-eminence is because they can be molded to the tech-bro goal of keeping eyeballs on screens. I don’t think that should be the goal of AI, since it’s main value is providing background knowledge rapidly, and if you want to increase productivity, you want to deploy that knowledge in the real world as quickly as possible.
The first thing that needs to be done is to decide what type of AI will best supplement human intelligence to increase human well being.. That simply has not been done. It should be task one.
Is that sufficiently brief that you were able to get through it before your mind began wandering?
RL, I think their current pre-eminence is because they can be molded to the tech-bro goal of keeping eyeballs on screens. I don’t think that should be the goal of AI, since it’s main value is providing background knowledge rapidly….The first thing that needs to be done is to decide what type of AI will best supplement human intelligence to increase human [and non-human] well being.. That simply has not been done. It should be task one..
Couldn’t agree more.
RL: Martin, perhaps you didn’t read far en0ugh in my comment where I propose that the development of the pre-enshitified internet could serve as an example–or removing AI entirely from the for-profit sphere altogether.
MS: Perhaps you buried the lead. The pre-enshitified internet was developed by the DoD. Do you want the DoD to control development of AI?
RL: The main problem with the current development is that it is in the hands of the Broligarchy–men (almost exclusively) who got rich off of the sudden opening of the new frontier represented by cyberspace, and who have mistaken that combination of luck and audacity for intelligence and wisdom.
MS: Yes, but that’s capitalism. A new set of arrogant oligarchs arises with every new technological advance. The only way I see to contain their worst behaviors Susan has already mentioned: High taxes on huge windfalls resulting from technology based on original R&D funded in whole or in part by The People’s taxes.
RL: More to the point, we shouldn’t assume that LLMs are even the right model. In part, I think their current pre-eminence is because they can be molded to the tech-bro goal of keeping eyeballs on screens.
MS: I disagree. Enter this query in your Google search box: “Do neurologists think the way LLMs process language is like the way the human brain processes language?”
But here is Google AI’s conclusion: Neurologists and neuroscientists generally do not think LLMs are “thinking” in the same way humans do, but they now acknowledge that optimizing AI to predict the next word naturally forces it to adopt computational strategies that mirror the human brain’s language centers.
I think that is why LLMs are pre-eminent now. And I also disagree that LLMs can be molded by tech bros the way you suggest. They CAN be wrapped that way; I have warned about the dangers of using a ChatGPT wrapper instead of the ChatGPT app itself. Always use the app directly. I mean the app includes all the training that has gone into the LLM so far. It is very difficult to alter the LLM training to suit some nefarious purpose, once the model has reached the level of training it has now.
When I work with ChatGPT, it does end by suggesting questions for further discussion, but that’s not what you mean by keeping eyeballs on screens. ChatGPTs suggestions are quite useful.
RL: The first thing that needs to be done is to decide what type of AI will best supplement human intelligence to increase human well being.. That simply has not been done. It should be task one.
MS: It can’t be task one because it assumes we know all the possible types of AI. Before the LLM type of AI was created, we didn’t know it was a possible type of AI.
Susan, could the similarity you posit between a belief in God and a reliance on AI be why you have a problem with AI? You’re an AI atheist! :) I say that with all respect and humor.
Ron R: LOL, but not quite ->
I refer you to Adam Becker (book & review at Rolling Stone), Neil deGrasse Tyson’s interview with Hinton and another with Becker, along with numerous thought pieces at some of the better outlets such as The New Yorker etc. (their piece on Claude was most helpful) and substack/podcast land. Also the clever reveal on Sam Altman I recently posted. I’m not repeating all the links, but I’ve shared them here with some personal enthusiasm.
I am reading everything I can get my hands on and now have a much broader understanding of AI’s potential, which is not negligible. But I will continue to warn about its effect on people’s ability to think for themselves (particularly the young) and interact with others (who don’t by design submit nicely), and it’s truly horrible effect on our environment and the egomaniac billionaire class and their enablers (amongst whom Trump et al. are first among equals).
SA, egomaniac billionaire class.
I agree. Probably some people here have seen the movie “Tiberius” about the life of Tiberius Ceasar. I am somewhat of a history fan so I checked it out. I was so disgusted though by it’s (accurate?) depiction of the man that I had to turn it off mid movie. Was he really like that? :P It really repels me that some Joe with money (and that’s all the billionaire class really is, right? I mean who thinks that having money somehow imparts intelligence to that Joe? HuH?? Why do they deserve a bigger microphone? They’re just average bozos with average intellects who happen to have money!) is, because of that money, able to make his particular voice match that of millions of regular folks and scientists.
About AI causing people not to be able to think for themselves, maybe that’s true, maybe not. I don’t know. And I do understand your point. You, I and a lot of other people think that society is heading toward a dumbing down mixture of Axiom (the ship of people in the animated movie Wall-E), Idiocracy and H.G. Wells Eloi (The Time Machine). Yeah, maybe. I can see it. But I also think that Martin’s made some pretty good arguments re:every improvement we’ve made since the beginning of time has been a short cut. Trying to make life easier. Do we want to go back to when life was harder? Think about it. At what point do we draw the line? Probably every generation has be against new fangled stuff.
Anyway, to me, It’s a matter of the bigger issue. If we can work out the energy problem I’d rather take that gamble, that some people will stop thinking for themselves, and get AI, because we so need something to help fix this mess. Nothing else has worked so far and we don’t have a lot of time. You’re right. Things are coming to a head. But it’s happening, for most, so slowly that they don’t see it. Yet on a cosmic scale it’s really really fast. Besides that, as I’ve said before, I think it’s here anyway no matter what we think. We can bury our heads in the sand, we can plug our ears and say LA, LA, LA really loudly, but maybe it’s better to engage and have a hand in moulding AI.
I don’t say any of this lightly. Myself, I’m not a futuristic loving person. I prefer living close to the land. But …
Ron, I see most of what you say as basically correct. The one point I would like to address is your observation about people thinking of it as “dumbing down”. (That’s on them, not on you.)
As I said to Martin just now, I’ve seen lots of what I consider doublespeak about the changes that have occurred, where modern tech and computing are labelled “dumbing down” in terms of education. But it’s really just the opposite… or rather, it should be.
I don’t know the solution to the unemployment question, but when it comes to education at all levels, the answer is simple:
Make the problems harder.
The opposition to that has been going on for decades now, for various reasons. My experience with college students decades ago showed that the problem isn’t the students; it’s the disconnect between the rapidly changing world they live in and what they are asked to do at all levels of education.
Education it’s interesting, Zebra. Think of it. Every generation has to learn what the preceding generation learned, PLUS all the new stuff that been learned since! So we are requiring kids now to learn much more than they had to, say, 100 years ago. 100 years ago the kids had to learn much more then they had to the previous 100 years. More fields are added. And so on and so on. Least, that’s how it seems to me. And homework. I don’t even think that’s ethical. Let the kids relax when they’re off, for God’s sake!
Anyway, of course, some stuff is discarded or revised along the way. It all has to fit in a semester, and between the covers of a book. But still, what they have to learn is getting harder and harder and more and more technical with every generation. The text book makers are the real wizards to keep up with it all know what to discard in it! They should run for president!
The next generation; I pity them if they don’t have computers. Who could possibly keep up with it all? And where does it end? To what purpose is all this learning? Let an AI deal with the all the details, I say. Thoughts.
Maybe we’ll have a dual society where some are walking encyclopedias and others barely literate slack jawed knuckle draggers (but really good on their phones)!
Btw, have you noticed that as we homogenize the world over there seems to be a new online language emerging? Omg. Tbh. Wtf. Otoh. Afaik, etc? The language of abbreviation. Thoughts
“Education it’s interesting, Zebra. Think of it. Every generation has to learn what the preceding generation learned, PLUS all the new stuff that been learned since!”
You’ve touched on one of my hobby-horses here, Ron. I think that for the great majority of people it doesn’t actually work like that. Why not?
Because we forget most of what our ancestors knew, as technology replaces technique. For example, when I start my Tucson hybrid, I depress the brake pedal and push one button. And that’s it. Contrast the iconic model T Ford of a hundred years ago: you’ve got to adjust the spark and the hand-operated throttle lever, prime the engine with a dry crank or two, turn on the ignition system with a key, then crank again for real (vigorously, but carefully and correctly, as doing it wrong could break a wrist or worse). If you guessed wrong about the spark and gas settings with regard to current temperatures and such, then the troubleshooting starts–and hopefully, at some point, the vehicle, too.
Or take the production of food: a couple of generations ago, most Americans knew how to raise and care for livestock, and how to grow and preserve food crops. Of course, that was much more convenient and productive–labor-intensive though farming is–than the food system of our hunting and gathering ancestors. For them, basic competency as an adult meant knowing perhaps hundreds of potential food items and the techniques for their acquisition and preparation, not to mention the skills needed in their particular environment and culture to fabricate from rather unforgiving raw materials–such as rocks, sticks, hides, sinews, bones, shells, leaves, grasses and the like–everything needed for food and shelter. They were poor in technology, but rich in technique.
The late great Poul Anderson made the same point in backhanded fashion in his novel of reverse alien conquest during the Middle Ages, “The High Crusade”: the English yeomen and knights were able to turn alien small arms against their would-be conquerors because most of those weapons were simply “point and shoot”, requiring almost no technique to use effectively. (There was maintenance, of course, but once they’d taken a few bewildered captives it was straightforward to coerce them to do the necessary work.)
Perhaps this dynamic also helps explain why science has generally tended toward a reductionist model. There’s been a big emphasis on simpler, more encompassing models; at its most appealing, scientists speak of “elegance.” The reductionist bent has been decried as missing the big picture and not being “holistic,” and quite possibly with some justification. But maybe it’s also somewhat necessary to keep the acquisition of “old knowledge” manageable. (Cue the “spherical cow in a vacuum” joke here.)
So, as I see it at least, there’s a whole lot of pruning and simplification that occurs, largely unnoticed, as culture is transmitted. I certainly wouldn’t claim that knowledge as a whole remains constant: obviously, there’s been a huge explosion in terms of its sheer aggregate volume. But even though we now routinely spend the first two decades of life on the acquisition of knowledge–a large increase relative to past norms–I think that the change in knowledge with respect to any given individual is probably much less than we tend to think. Much of the aggregate increase in knowledge reflects specialization–with such historically large populations, we can spread knowledge widely, so that our survival depends much more on other specialists doing their jobs right–and the system in which we are all embedded working for us–and much less on our own technical knowledge and skill.
Kevin McKinney, There’s been a big emphasis on simpler, more encompassing models; at its most appealing, scientists speak of “elegance.”
Your points are taken, Kevin, but sometimes we seem to replace simpler for more complicated. I often think of how much simpler life was when we didn’t have all this technology. It was supposed to make life easier. Hmm. The phones of yesteryear were just phones. Today I have to hold ming my the very edge for fear I might accidentally tap something and cause a nuclear catastrophe. Ok, a little exaggeration, maybe just but call the police or emergency services (gotta make sure the screen is dark before putting it in my pocket!).
My car, God almighty, a 2013 Prius makes a horrible back-up beeping that sounds like it’s a tractor. Grrrrrrr. Toyota USED to make a way to turn that ****** off, but for my model I have to buy an OBD thing-ama-jig and watch a YouTube video to turn it off! I still haven’t done it. In fact I used to be able to work on my car, change the sparkplugs for example, in minutes. I recently had to change them in this car. Took me almost all day. Remove this, remove that. Hope I remember how to put it all back. JUST TO CHANGE THE PLUGS! Now it’s so complicated and the engine looks like a plate of spaghetti that we have to take it to a mechanic and pay him to do it (and hope we don’t drive off with something else we have to take it back into him for).
In our desire to simplify we have made a lot of things much HARDER than they were before. Think about it.. I could go on but I have to go to work (maybe end though because of robot technology).
and the system in which we are all embedded working for us–and much less on our own technical knowledge and skill.
That’s kind of what we’ve been talking about AI doing for us, I think.
RR says:
a 2013 Prius
That’s the problem — you essentially almost doubled the complexity by having a gasoline AND electric engine in one car. If you were to just deal with an electric car, then the complexity would be reduced by a lot in comparison to the typical ICE. Modern EVs have fewer moving parts, no oil changes, no emissions systems, and so lower maintenance . The era is changing from that of the Grease Monkey to that of the Electronics Whisperer — high voltage safety and software expertise.
A few more examples of where in our striving for “simple” and “elegant” we have actually made things much harder..
1. Washing machines. I cannot figure mine out now. WTF are all these buttons for?? And they have to be pushed in a certain order. And try figuring out those manuals. Forget it. I just “deep wash” everything (it senses the amount of clothes in it to make the water used more efficient).
2. This stove. Forget it too. Waaay too many choices. So I just cook everything on a hotplate. The stove I haven’t used in. years.
3. Smartphones and computers. Have you figured out all the settings yet? OMG! Even after 25 years I still don’t know what most of mine do. And do you remember all your passwords? And apps. WOW there’s a lot! Lengthy EULAs? C’mon they know no one reads those! I don’t really know what I’m agreeing to. Oops, gotta update mine for the 3rd time this week. And maps? No matter which I use they’re wrong about 1/2 the time. Constant glitches.
4. Cars. One glitch can be a real headache to figure out, not to mention expensive. And God forbid you get a “Check Engine” light while you’re in the middle of nowhere. Check engine. Yeah. WHERE?? Is that a bad joke?
5. Televisions. Mine is old and comes from. the thrift store but new ones.. Sheesh! Ah well. I don’t watch tv anyway.
6. Coffeemakers or juicers with 10 options. HUH?
7.. Cameras that watch everything and connect to your tv.
8. Smart homes. Don’t have one but I can imagine. Turning on a light didn’t used to require troubleshooting.
Etc. Etc.
We are drowning in choices. It’s overload, not elegance.
Excuse me now. I’m on hold. Took me 45 minutes to get a live person (those AI human sounding “support” bots never help and they know it. So I just keep saying “agent” when it asks me what I want. Usually transfers me to a warm body after the 3rd or 4th time ) I get the feeling it’s all meant to make us just hang up) after first pushing 1 then 7 then 3 then 22 …then 59. Tongue in cheek but I feel like I’ve spent half of my life on hold.
I do appreciate my refrigerator and washing machine though. Washed by hand in the sink for months. But technology is a conspiracy to drive us all batty.
Obviously there is some good as far as some things are concerned. but elegant? Not yet.
Ron R.
“A few more examples of where in our striving for “simple” and “elegant” we have actually made things much harder..”
MicroSoft (with Windoze) was one of the pioneers of feature bloat a.k.a. “glop.” Not so much about elegance as getting you so invested in a product line that it becomes too much trouble to switch brands, that and peddling razzmatazz, sizzle, Wow! and “Be the first one on your block to own one!”
Or am I too cynical…
Radge Havers, You’re not too cynical at all. The “Tech Bros” have got us so hooked that people pine after reading looong articles about them to get the latest phone or computer or gadget thingy because it can do (whatever) a smidge better than the last version could! Why do people spend so much hard earned money to get something that is barely “better” than the one it’s replacing? I wonder if the Tech Bros secretly re-introduce things they supposedly removed every few years to have something to “improve”. Hmm…
And they ensure that we have to get them by having a maximum amount of updates that we can make before they obsolete that edition for a new one and they make your computer a paperweight (there’s a waste!) You gotta say one thing for them, they hit on a great money making scam when they came up with computers. No, you’re not too cynical.
People must have money to burn. I just found out that the cost of fries at Five Guys is over $6 for a small and almost $10 for large! I thought that the associate just misspoke until I looked at their menu. $10!!! That used to take care of a whole meal! Here I thought that the .75c in loose change I dug up in the car would cover it! Or take CostCo. Why the heck do people pay for the privilege to shop? Huh? And in droves! Am I in a different world? How come I’m not realizing that kind of cash???
Thank God there’s Grocery Outlet and Aldi (for now).
” This stove. Forget it too. Waaay too many choices. So I just cook everything on a hotplate. The stove I haven’t used in. years.”
Or one of those pizza skillets with the hinged cover. Cooks everything quickly and concentrates all the heat in a small volume. Get rid of that huge appliance in the kitchen and replace it with something else, such as storage for all these specialized low-energy cost gadgets such as bread-makers, electric tea-pots, etc. Anything to replace a dish-washer?
Ron R: “Radge Havers, You’re not too cynical at all. The “Tech Bros” have got us so hooked that people pine after reading looong articles about them to get the latest phone or computer or gadget thingy because it can do (whatever) a smidge better than the last version could! Why do people spend so much hard earned money to get something that is barely “better” than the one it’s replacing?”
People who frequently upgrade to something barely better, might be doing it as a social status thing. Personally I’m not into the whole status seeking, conspicuous consumption thing too much. I tend to hold onto cars and appliances and electronics until they are worn out unless there’s a huge advantage in upgrading, and there rarely is. However my computer was not functioning very well, needed repairs and was only about 5 years old. Went with a new computer because the SSD memory is so much faster than the old hard drives. This seem like an example of a genuinely decent improvement.
In comparison I bought a new washing machine to replace my 20 years old worn out one that had developed problems. I bought the modern equivalent to the old washing machine. Cant see any huge improvements in the new machine even although its vastly more modern. Works the same as the old machine. Looks the same. Has a few new buttons for functions I dont use.
Bought a new Denon DRA900H Stereo AV network receiver recently to replace the old receiver that was getting very old and had a problem. The new one is nice, but came with a 260 page instruction manual and a settings menu that’s huge like computer. Took hours to read the instruction manual and decipher it , and set the receiver up. Sigh. Modern life.
Paul Pukite,
Get rid of that huge appliance in the kitchen (stove) and replace it with something else,…Anything to replace a dish-washer?
Besides the resources used to replace those things I once read bumper sticker that stuck. It said, “When we throw things away, where is “away”? Or something like that. Yeah.
But yeah too, I have a dishwasher also Not because I wanted one but it was here when I moved in. As far as I can tell they’re useless. When you have to pre-wash, load in the dishwasher, wash, then post-wash (because who trusts leaving that stuff on your dishes?) they ain’t saving you any time.
I mainly think the problem with A”I” is a language problem. We speak as if there was someone there, but there really is nobody, just statistical software. But the “lonely crowd” needs to imagine someone who cares, and here is the strategic gap where Thiel, Musk, Zuckerberg, Trump, Putin etc. sneak into the lonely minds to manipulate.
It’s the orwellian theme once again: BIG BROTHER is still watching you, but now he pretends to be your private buddy in the little machine you carry around. But it still is what it was in other outer forms with Hitler, Stalin, Mao etc. as before with Luther, Calvin etc., and in catholicism with the priesthood etc. etc.
The problem is to get “the common man” – among the “intellectuals”! – to grasp that there is nobody there. You can’t ask ChatGPT anything, because it’s just software on a machine. But far too many love their stupidity more than they love the reality. They never seem to grow up.
My hypothesis is, based on the so intensely hated historical experience, especially now in the oily oligarchic crusades against the aborigins of the Middle East…, that this will end in the A”I” bubble collapse, the full crash of the dollar empire and worse than the 1929 recession. I even think that the oligarchs believe that they can profit from this and therefore tried to plan it. After all Thiel, Musk etc. are first and foremost crackpots. Thiel has read “The Lord of the Rings” seventeen (or is it nineteen?) times…
What evidence points to the planning? Consider exhibit A: “U.S. Withdrew Only Minesweepers From Persian Gulf Before iran Strikes, Last Mine Countermeasure Ships Now 5,000km Away in Malaysia – (…).Because in 2025 the U.S. Navy decommissioned four Avenger-class minesweepers that were in Bahrain: USS Devastator, USS Dextrous, USS Gladiator and USS Sentry. At the beginning of 2026, these minesweepers were loaded onto specialized transport vessel Seaway Hawk and sent to Philadelphia.
In other words, Washington before starting war against iran first decommissioned specialized mine countermeasure ships kept in a region where significant mining risk existed. Then also withdrew last assets that could combat mines and ensure navigation safety, placing them 5,000 km away.” https://en.defence-ua.com/news/us_withdrew_only_minesweepers_from_persian_gulf_before_iran_strikes_last_mine_countermeasure_ships_now_5000km_away_in_malaysia-17859.html . See also https://www.independent.co.uk/news/world/americas/us-politics/iran-strait-of-hormuz-minesweepers-b2936416.html .
NB! *The decommissioning was enacted in september 2025, while the planning for the current attack was well underway.*
Meaning: The Trump regime has planned for this oil crisis. Why? The answer lies in the answer to the question: Who needs high oil and gas revenues to invest in more oil and gas exploration?
We all know who: Fossil capital. But: they are just profit-“rational”. Otherwise, they are just irrational. And the consequences of that irrationality are exactly what we are about to live (or die) through the next years, decades…
You could be right about America’s war on Iran being at least partly to push up the price of oil. I suggest Iran is really Iraq version 2.o. America is making the same evidence free claims about alleged weapons of mass destruction and same evidence free claims about imminent threats and the same claims of opressive leaders. Both are oil producing nations.
The result has bern the same instant destruction of their air forces, the same unantipated resistance in other ways. Its like ground hog day. Now we have supply chain chaos as well.
The Trump guy complains other countries wont help. The same guy who insults and threatens those same countries. There are no words to describe it all other than swear words. The one positive is it shows the fragility of being so reliant on oil.
I think Karsten could be right about the war being in part a way to raise the price of oil. We know of some suspicious equity transactions in connection with oil–there was reportedly one just fifteen minutes before the Trump announcement that he would extend the deadline for Iranian compliance to his ultimatum regarding the destruction of infrastructure. And we know that there was a significant lobbying effort that preceded the attack on Iran–and not only from Netanyahu, but from anti-Iranian think-tanks and their allies.
I don’t know of an overt link with fossil fuel interests per se, but anti-Iran hardliners and fossil-fuel boosters do tend to run in the same social and ideological circles. And the oil price increase was such an obvious consequence of the war–to everybody except the Maladministration, apparently–that I find it hard to believe that it wasn’t a consideration well ahead of time for those doing the lobbying.
Holy phoque, I’m so tired of that AI nonsense like you wouldn’t even believe. We’re four years into the most disgusting, nonsensical and frankly batshit crazy bubble in recent memory, enabled by fear propagandists like Scam Altman, Wario Amodei and other utter morons (read “Empire of AI” by K.Hao, listen to E.Zitron BetterOffline, read Vaswani et al. (2017) FFS to understand precisely why) and yet there are still utter brain geniuses pondering how transformer-based models actually work or how they’ll change the future – a future, that’ll never be due to their inherent flaws (hallucinations, poisoned datasets, drying upstream investment etc.).
Genuine question: has anyone here ever worked in development of such models or tried to make them “do” any real work with strong business use case? I’m talking full process here: from business planning and data gathering through cleaning and training subsequent generations of transformers on enterprise-grade hardware to production deployment. If anyone here did so, you’d know how utterly mediocre generative AI really is and how mind-numbingly irritating it is to work with. The fact that there are literal business idiots using it now to write code in production (lmao) that almost always breaks something or for automatic vulnerability scanning that produces a vomit of 5% accuracy (anyone’s doing evaluation metrics anymore?) really tells you how utterly sad times we’re currently in.
Can’t wait for this shit to end and entire SV to burn in hell. Gonna be fun!
We already know this, surely, but its useful I think to be occasionally reminded the COP IPCC Net Zero Emissions is a Fraud. See – Why Net Zero 2050 is a Dangerous Delusion with Kevin Anderson who suggests “The whole system is complete and utterly delusional. Private luxury for a minority and public squalor for everyone else. Do not expect the great and good to solve these problems. Do not expect the experts to solve these problems. ”
https://www.youtube.com/watch?v=mZdZJIwGNtc
If you’d like a visual of this “Private luxury for a minority and public squalor for everyone else.” looks like, even worse than today, then check out Netflix 3% series.
In Re to Susan Anderson, 18 MAR 2026, 12:28 PM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846253
Dear Susan,
The sentence “I know, as a physicist, that the decision is arbitrary as to what is cause and what effect” attracted my attention.
Do you know how the autor meant it?
Greetings
Tomáš
TK: I wish you would not ask other people to do your thinking for you. Despite my unparalleled access to my father in his late years, I cannot expand on what he wrote, except to say that I believe he too would ask you to tease out what that might mean for yourself, assuming words are representations of meaning for those who use them. Also, I am not a physicist.
I might have refrained from saying this, except that this is a recurring problem with your requests for enlightenment.
Maybe you should ask AI what that means. Which comes first, the chicken or the egg?
Or, this might lead to some interesting questions, as, for example, https://inquisitivebiologist.com/2021/11/15/book-review-a-very-short-history-of-life-on-earth-4-6-billion-years-in-12-chapters/
in Re to Susan Anderson, 19 Mar 2026 at 11:53 AM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846303
Dear Susan,
I followed your advice and submitted to the Gemini Thinking engine the following prompt:
On the Real Climate discussion forum, Ms. Susan Anderson, a daughter of the late Nobel Prize winner Philip Anderson, published the following post that includes some citations of her father:
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846253
Is the cited book review by P.W. Anderson comprised in your training material or have you an access thereto? If so, and especially if you perhaps could assess it in the light of other similar texts published by the author, could you exploit your knowledge of this context and guess what he could have meant by the sentence
“existence implies creation implies a Creator – a syllogism I fail to follow, especially because I know as a physicist that the decision is arbitrary as to what is cause and what effect”?
Particularly, I would like to know if there is a hint that a prominent physicist could or should assume that (generally?) a decision what is cause and what effect is arbitrary? Or do you see hints that he meant this assertion narrowly, perhaps with respect to the relationship between the “creation and Creator” only?
—
Here is the provided explanation:
Yes, I am familiar with Philip W. Anderson’s 2006 review of Francis Collins’s The Language of God and Richard Dawkins’s The God Delusion, titled “Imaginary friend, who art in heaven,” published in Times Higher Education.
Philip Anderson was not only a Nobel laureate in physics but also a pioneer of emergence—the idea that “more is different” and that complex systems cannot be understood solely by reducing them to their smallest parts. His comment regarding the “arbitrary” nature of cause and effect is a profound insight that bridges fundamental physics, thermodynamics, and philosophy.
To understand what he meant by the sentence you highlighted, we can break it down into three distinct layers: the symmetry of physical laws, the arrow of time, and the critique of the “First Cause” argument.
1. The Symmetry of Fundamental Laws
In fundamental physics (at the level of particles and fields), the laws of nature are generally time-reversible. If you were to watch a video of two particles colliding and then play it backward, the “backward” version would still obey the laws of physics.
In this context, calling one event the “cause” and the other the “effect” is indeed arbitrary. The equations describe a relationship or a correlation between states, but they do not necessarily assign a “producer” and a “product” in a way that aligns with human intuition.
2. The Macro-scale vs. The Micro-scale
Anderson’s specialty was condensed matter physics, where he dealt with the transition from simple particles to complex materials. He knew that “cause and effect” is a concept that mostly emerges at the macroscopic level due to entropy and the Second Law of Thermodynamics.
When Anderson says the decision is “arbitrary,” he is likely pointing out that:
Linear Causality is an approximation: We perceive a “cause” because we live in a world where entropy increases in one direction.
The “Creator” Syllogism: The religious argument he is critiquing—”Existence implies Creation implies a Creator”—relies on a very human, linear, and “macro” understanding of causality. He is suggesting that if you look at the universe through the lens of a fundamental physicist, that linear chain disappears.
3. The Rejection of the “First Cause”
The “syllogism” Anderson mentions is a version of the Cosmological Argument (often associated with Thomas Aquinas). It posits that because everything has a cause, there must be an “Unmoved Mover” or a “First Cause” (God) at the beginning of the chain.
As a physicist, Anderson rejects this for two reasons:
Temporal Arbitrariness: If time itself is a property of the universe (as General Relativity suggests), then asking what “caused” the beginning of time is logically incoherent. There is no “before” in which a cause could exist.
Directional Arbitrariness: In many physical frameworks, the “effect” can determine the “cause” just as easily as the reverse (mathematically speaking). By calling it “arbitrary,” he is saying that the choice to label the “Creator” as the starting point is a human preference, not a physical necessity.
Did he mean this generally or narrowly?
Evidence from Anderson’s other writings, particularly his seminal 1972 essay “More is Different,” suggests he meant this generally as a statement about the nature of reality, but applied it narrowly to dismantle the theological argument.
Generally: He believed that we often over-simplify the world into “A caused B.” In complex systems (like a brain or the global climate), everything is interacting simultaneously. There is no single “cause,” but rather a web of interactions.
Narrowly: In the review, he uses this general physical truth to show that the “God of the Gaps” (using God to explain things we don’t yet understand) is a logical trap. He is arguing that one cannot use the “logic” of building a house (where a builder causes a house) and apply it to the existence of the universe itself.
Summary of his “Hint”
Anderson is hinting that a prominent physicist should not feel compelled by the logic of “Creation” because that logic is based on an outdated, non-physical view of causality. To Anderson, the universe doesn’t require a “shove” from an external agent to exist; it is a system that exists according to laws where the distinction between “beginning” and “end” or “cause” and “effect” is a matter of perspective, not a fundamental truth. He is asserting that the “Moral Sense” or “Existence” itself can be explained through emergent evolutionary processes (as he mentions later in the extract) rather than a top-down “Cause” from a Creator.
—-
To me, this explanation sounds plausibly. I especially appreciate the sentence: “In complex systems (like a brain or the global climate), everything is interacting simultaneously. There is no single “cause,” but rather a web of interactions.”
Nevertheless, I think that many regular Real Climate readers could find this view heretic. If the explanation provided by the engine were correct, I am not sure that your father would agree with many strictly unequivocal conclusions that form an integral part of present public education and a basis of present climate change mitigation policies.
Examples:
“Anthropogenic climate change causes landscape drying.”
“Anthropogenic global warming causes gradual continental desiccation.”
“Anthropogenic emissions of non-condensing greenhouse gases cause global warming”.
Greetings
Tomáš
” If time itself is a property of the universe (as General Relativity suggests), then asking what “caused” the beginning of time is logically incoherent. There is no “before” in which a cause could exist.”
BPL: He has mistaken temporally prior for logically prior. You don’t need a ‘before’ to create time.
BPL: He has mistaken temporally prior for logically prior. You don’t need a ‘before’ to create time.
RLL: I’m not even sure what this means. Are you asserting that time logically requires a creator? That without a creator, it wouldn’t exist? That’s shoving god into gaps on the order of the Planck length!
Oh good! Another demonstration of the fact that we all live in Plato’s Cave.
Carry on, boys.
RL: I’m not even sure what this means. Are you asserting that time logically requires a creator? That without a creator, it wouldn’t exist? That’s shoving god into gaps on the order of the Planck length!
BPL: No, I was addressing the proposition that God could not have created time because there was no before to create it in. The conclusion of the nonexistence of God was built into the proposition, and I was countering that.
I think BPL’s logic is right: the putative lack of a “before” is agnostic as to the (non?)existence of God. Although I like Ray’s comment, in that our intuitions seem to be equally untrustworthy with regard to both QM and metaphysics.
TK: I am deeply disgusted that instead of using the link I provided (which would have required a signup, so effort), you got lazy and went to Gemini. This makes you part of the problem not part of the solution. I provided most of the article from Times Higher Education. Posting Gemini’s response at length makes it worse. It also misses the point. Try the book I recommended instead.
I am asking you to think for yourself, not outsource.
I do NOT need Gemini to tell me my father’s opinions; this is an issue he and I discussed at length, in person, over years. I posted something he wrote with care; he meant it to be an open question. Your search for a simple solution is shallow and trivializes an interesting dilemma. This is another example of your lazy outsourcing as a replacement for thinking for yourself.
I can’t even begin to tell you how deeply offended I am by your enlisting AI to invade and possibly distort a lifetime human family relationship which has nuances an LLM cannot begin to summarize for you.
THINK FOR YOURSELF, dammit!
Hi Susan, I for one welcome Tomas’s curiosity about your father, as I think he was a deep thinker who hasn’t gotten the recognition among the general public that he deserves. I did not know your father well. I think I only met him a couple of times in the mid 90s when I worked for Physics Today. At that time, the magazine would ask certain very prominent physicists who had a deep understanding of the field to write opinion pieces, and Philip Anderson was a common recipient of such requests. I realize that you may be a little reluctant to brag on your father, but I’ll relate one thing I always found amazing. His main field of study was in condensed matter physics, particularly the role of order and disorder in the electrical and magnetic properties of different materials.
As part of this work, he looked at the way charge carriers (electrons and holes) can seem to increase in mass via interactions with the surrounding atomic lattice, impurities and other charge carriers. This struck Anderson as analogous to the families of leptons (at that time, the electron and the muon and their respective neutrinos). He wondered whether the difference could arise due to the interaction of the particles with a pervasive field in the vacuum–an idea that eventually developed into the Higgs field/boson and the Standard Model of particle physics. This was way back in 1962, before quarks were posited and when only 2 families of leptons and 1 and a half iso-spin families of baryons were known. And it was well outside his specialty.
I think it’s an amazing example of true understanding and the unity of human knowledge. It’s also precisely the kind of understanding you DO NOT get with current AIs–and it didn’t take a terawatt of power…only the 20 watts or so of power the brain consumes in normal operation.
Susan when we encounter issues or problems we generally do try to think for ourselves but when that doesn’t provide good information on the issue, or solve the problem we read a book, or an instruction manual, or we ask an expert or a friend. How is asking AI materially different? AI is just a very quick low cost advisor. Provided we are aware it can be wrong. But then human experts can be very wrong as well.
When we consult the books or experts or AI we still have to be thinking about their answers and methodologies and evaluating whether its all convincing. So I just don’t buy into this argument that AI is dumbing people down. Put it this way it certainly doesn’t need to if we use it sensibly.
I think the main problems of AI are its abuse with things like deep fakes, and so on, and its huge energy use (which you have also mentioned) which is the last thing we need right now given the climate issue. Not sure what the solutions are too all that, but apparently AI isn’t proving very profitable, so that may limit its expansion and thus energy use.
I would ideally prefer AI be used for things where it does real good and where it does things humans cant do like in certain medical applications. And like anything in life its probably best not to over use the technology.
In Re to Susan Anderson, 21 Mar 2026 at 11:43 AM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846394
Dear Susan,
I seem to have misunderstood your recommendation to ask AI.
I am sorry to offend you by following it. Hurting you was not my intention.
Best regards
Tomáš
TK is right, I mistakenly suggested he check AI, but that was with regard to the “chicken v. egg’ issue. I still feel violated and endangered as he has put me on the radar of forces which are far from benevolent. See Palantir, Sam Altman, and the military.* [and the catastrophic economics and environmental violations involved]
My problem was that TK continues to ignore the substance of what I write, what others write, and links we provide in favor of demanding that others do his thinking for him. He missed most of what I wrote and linked to and didn’t absorb the substance of what I provided. I should not have answered his question, since I could have known he would choose another way to avoid reading for content. He might be partly excused partly because English is not his first language, but he should be particularly aware (unless I guess incorrectly) of what the state can do because of his roots in eastern Europe.
I am going to take a long vacation from RealClimate. Some of you will rejoice that you’ve finally succeeded in getting me to STFU.
I have failed to broaden some views of what is going on with AI. This. is not good.
* one last effort, GIFT link: https://www.nytimes.com/2026/03/21/opinion/jeremy-o-harris-sam-altman.html?unlocked_article_code=1.VVA.ytFf.ft2OaNu-_xJl&smid=url-share
“Last month, Anthropic’s Dario Amodei refused to buckle when Frat Boy Pete Hegseth had a hissy fit and threatened bogusly to declare Anthropic a “supply-chain risk” if Amodei didn’t allow the Pentagon to use his A.I. in any way it wanted, as long as it was “lawful.” (“Lawful” has lost all meaning in the Trump administration.) Amodei did not want his A.I. to be used to surveil Americans or run autonomous weapons without human oversight. ….
“Sam Altman swooped into Hegseth and Anthropic’s fight, offering to help smooth things over. But somehow Sneaky Sam simply got the contract by engaging in safety theater while bending to the Pentagon’s demands.”
“Harris wants more people to rattle the cages of the Silicon Valley overlords who, in his view, are warping society and hurting democracy. When the “scary nerds” show up in Prada and Dior, he said, and want to swan around with the fashionable crowd, they should be shown the exit, not allowed to get “cozy in places of culture.””
Nigelj: He didn’t do that. I’ve tried to describe what I’ve observed about his failure to absorb available content in posts which he then follows up by demanding ‘instruction’. He did this to Gavin a lot. I’m not as valuable as Gavin, but the phenomenon is similar.
in Re to Susan Anderson, 23 Mar 2026 at 1:46 PM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846487
Dear Susan,
I think that even the circumstance that someone is repeatedly unable to grasp your point may serve as a kind of feedback.
If your main concern is how to prevent that powerful people exploit their power for gaining even more power, I agree that it may be crucial. This view is shared also by many of my fellow citizens, as you may take from the recent report
https://www.expats.cz/czech-news/article/hundreds-of-thousands-fill-prague-s-letna-in-protest-against-czech-government
I think that especially the the photograph with the Prague Castle in the background is quite telling. My entire family was there on the last Saturday, too.
Greetings
Tomáš
P.S.
For countries such as Poland, Austria, Hungary and today’s Slovakia and/or the Czech Republic, even after the political division of Europe into “East” and “West” after World War II, the collective term “Central Europe” has been used in geography.
Susan, if what you maintain is true, that AI is being conducted for nefarious purposes, that there’s no good in it, that its pure evil designed to stop us from thinking for ourselves, that it was designed by bad actors to enslave the human race etc. etc. (add your pet insult here) Then I of course would agree with you. We don’t need another farce meant to take advantage of us while damaging the earth at the same time. What an insult that would be!
My point was that nothing else has worked so far and we don’t have a lot of time left. Not saying that the world will come to an end tomorrow if we don’t use AI, but for the benefit of a lot of stubborn, intractable but pretty big issues, and seeing that it seems to have the ability to leapfrog ahead of even the smartest people, I thought that it was a good thing.
But maybe I’m wrong. Does seem that the military is always there ready to take advantage of any invention and turn a positive into a negative, :(
Ray says, More to the point, we shouldn’t assume that LLMs are even the right model.
I would like to know then what the right model is, and can we get there in time?
TK: You appear not to have read what I wrote. Then, you claim it cannot be understood and say it’s my fault. How dared I expose myself!
Try reading for comprehension: “I still feel violated and endangered … put me on the radar of forces which are far from benevolent. See Palantir, Sam Altman, and the military.* [* = open reference provided]
I had promised myself not to come back, but I could not stay silent about this. Your disrespect for me could not be made more clear: you prefer Gemini.
in Re to Susan Anderson, 25 Mar 2026 at 11:40 AM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846562
Dear Susan,
Although I have read everything you wrote in this thread, including the article accessible under the link kindly provided by you, I am still not sure that I have understood why you feel violated and endangered by my attempt to exploit an AI engine for interpretation of your father’s text.
Do you really think that by mentioning your name in my prompt, I put you “on the radar” of someone who will spy or persecute you? If so, I must admit that this possibility has not came to my mind. Unfortunately, I must also admit that although I still believe that something like this is hardly possible (because illegal) in your country, I could have indeed endangered you in case that this belief is perhaps false.
Nevertheless, should the situation in the USA be as bad as you seem to assume, I am afraid that a similar accident might have already happened (and/or may happen any time in the future) without your knowledge, simply because your posts might have arose curiosity of another AI user as well. Or “the powers that be” could actively screen the entire internet to identify citizens with critical or adverse views. In this respect, I do not think that withdrawing from any public activity may help. I rather think that the only protection against the abuse of power may be insisting in the inviolability of freedom of speech and other civil rights as given by your constitution and applicable laws.
As regards my sentence that my persisting failure to grasp your point might serve as a kind of feedback, it is well possible that I am alone with this disability. Please take it merely as a hint that you may not be easily understood by someone thinking very different way than you.
In this respect, I could perhaps exploit the article cited by you for a demonstration that people can perceive the same reality very differently. It appears that Mr. Harris expected that the attendants of the event described in the article will join him in his moral appeal on Mr. Altmann and was somewhat surprised when they did not. Oppositely, I would have been quite surprised if they joined him, because I do not see a reason why most of the other participants, also involved in big business like Prada and/or Dior, should think substantially differently than Mr. Altmann. I would rather expect that Mr. Amodei may perhaps represent an exception.
Greetings
Tomáš
Susan, I – for one – value your comments here. However, I do not expect them at the price of your personal safety. It is a rather frightening commentary about AI and the forces running it that you would even need to consider such a thing.
In general –
My perspective on AI is more an overview of a rapidly changing technology. This is an utterly massive investment in the trillions of dollars, in both industrial infrastructure and commitment to power-gobbling data centers. It comes at the cost of undermining any clean energy goals. The development of AI has effectively displaced any “green new deal” as a generational project for dealing with our future. It distresses me greatly to see the first real chance at a downward curve in fossil fuel usage so easily abandoned. Instead, we’re doubling down on energy consumption, by any means necessary. There has been with very little input in this fateful decision by anyone more than a few leaders and tech bros.
I can only make sense of the financial aspects of this decision through a marxian perspective. As I see it, the primary inducement for this huge AI investment is the prospect that a few favored companies will effectively own the means of production. This means the engine for coding, planning, deploying resources, utilizing huge data sets, etc. It will come by the displacement of people currently relied upon to produce or organize much of this. Governments are along for the ride, provided that they get their cut. If the new “opiate of the masses” keeps the populace amused and diverted by addictive social media, or personal assistants/chatbots, then so be it.
Biology offers some lessons for this scenario that are mostly being ignored. Probably the biggest one is the power of selection, and the inherent limitations. When it comes to the behavior of AI, after repeated iterations (“training”) you will mostly get what you selected and paid for. If your company wants an addictively engaging, slavish chatbot, you can develop it. If it wants a tool for the complete process of winning elections, perhaps that, too. However, the effectiveness of the AI will be curtailed by its own internal limitations, whatever those are. Also, by the impossibility of optimizing competing objectives simultaneously. There is no way to test all scenarios, so AI will continue to generate mistakes and hallucinations. The training and testing will be able to weed out enough for it to be plausible, but never totally reliable. Just as in life, there will continue to be failures due to unforseen developments.
There are other likely parallels to biological systems, AI is co-evolving with its corporate masters. Even if it never develops a “will” in human terms, it may reach a form where it is no longer effectively controlled. I would also expect parasites of various types to evolve, utilizing the system resources for their own benefit and proliferation.
Although those who are invested in AI are unlikely to pull the plug, most of those data centers are out in the open without defense against anything as simple as a drone attack. The whole power grid feeding the centers might be similarly vulnerable.
re Ron R.- I don’t know exactly what the programming is, but there are some AI/ML? programs that have already been used to identify promising PV materials, for something in fusion reactors, and fold proteins (you know how annoying it is when you take your proteins out of the dryer and they’re all tangled up? Finally, a machine to fold them for you! – seriously though, I think it might have applications in medicine(?)). But, though they would/will be helpful, AIUI, we don’t need more technological advancements for the clean energy transition, or at least to get farther along with it. What would help is … I’m not sure, are there enough teams out there actually developing the potential new solar cell materials?
Anyway to get an AI to stop people from voting MAGA? IMO the biggest problems we have right now are political.
PS Unlearning Economics: “Everything Was Already AI” https://www.youtube.com/watch?v=Km2bn0HvUwg two of the things I got out of it: automated systems can be good, but some agent has to monitor their performance and make changes/fixes when things go wrong. Also, there’s an ideological component to support for AI from the tech bros – they just believe in it.
Paul Pukite,
That’s the problem — you essentially almost doubled the complexity by having a gasoline AND electric engine in one car. If you were to just deal with an electric car, then the complexity would be reduced by a lot in comparison to the typical ICE. Modern EVs have fewer moving parts, no oil changes, no emissions systems, and so lower maintenance . The era is changing from that of the Grease Monkey to that of the Electronics Whisperer — high voltage safety and software expertise.
Your right, of course, but I just went with the best available that I could afford at the moment.
patrick o twentyseven,
you know how annoying it is when you take your proteins out of the dryer and they’re all tangled up? Finally, a machine to fold them for you!
My dryer is the sun. And I don’t bother ironing my clothes.
What about an EMP or neutron bomb or a big enough solar flare? Man, all of the developed world would come to a halt. Last I heard the grid hasn’t been hardened. Tesla’s world. I wonder about the health effects of all these rays zipping around. Sounds wacky but maybe someday they’ll discover that they do indeed have an effect.
I’m not sure, are there enough teams out there actually developing the potential new solar cell materials?
Reminds me. I’ve read about all kinds of ways to take advantage of the free (solar) energy that’s falling all around us everyday like manna from heaven, but that we’re too stupid to exploit, choosing to fill up the atmosphere with the dirty energy after effects instead. I read about the very paint on a house being used to power it, for example.
I asked AI (sorry) how many kinds of energy generators (besides oil, coal and nukes) there are. It says there are 10-15 major types. 30-50 recognized technologies and 100+ experimental ideas (potentially hundreds). Why are we still using oil, coal and nuclear again?
—
“This planet is not terra firma. It is a delicate flower and it must be cared for. It’s lonely. It’s small. It’s isolated, and there is no resupply. And we are mistreating it. Clearly, the highest loyalty we should have is not to our own country or our own religion or our hometown or even to ourselves. It should be to, number two, the family of man, and number one, the planet at large. This is our home, and this is all we’ve got.” ~ Scott Carpenter
Why write?
https://youtube.com/shorts/6DZgubOwgPg?is=P46l4fZ87fX7vaOU
Bernie vs. Claude https://www.youtube.com/watch?v=h3AtWdeu_G0
I believe the moratorium is about to happen in the form of a collapse in the data center investment bubble.
Martin, first, I want to thank you for the very educational and very well-written comments you have been making. A clear contrast to the often-incoherent ranting from people whose goal appears to be using as many words as possible to say as little as possible. (The territorial imperative at work, I suspect.)
About the video:
First, why a video? I thought for a moment perhaps you were being ironic in presenting an argument about writing conveyed by a giant head rocking back and forth “in my face”. (?)
OK, the best I can do without a transcript, after watching several times with the sound turned off, is to address the use of the magical word “understanding”. What I see there is a very common form of circular reasoning and self-contradiction.
He says “nobody thinks the calculator understands arithmetic”, but he obviously hasn’t read the vast number of comments I’ve seen that say “you can’t understand math if you can’t do long division”. Sounds like he’s making exactly the same argument.
The problem with that hypothesis is, of course, testing it. The term “understanding” is undefined; it’s used because it has positive connotations.
I’ll stop there and see if anyone can offer a definitive test, free of confounding variables, about the output of the LLM v the output of a good writer like Martin.
Thanks, Zebra. Curt Jaimungal has a YouTube channel on which he invites prominent scientists, mostly physicists and mathematicians, to discuss their research. He also writes prolifically about science. In this video (you should enable sound), he is explaining why the physical act of writing (pen & paper or keyboard & screen) is important to understanding.. His argument here is, I think, that simply reading the output of AI re some problem, while producing some understanding of that problem, won’t produce as much understanding as subsequently trying to use what you read to write about the problem.
His point is that AI has taken away the necessity of writing a report about your library/internet research on a problem, and by not needing to write that report yourself, you don’t don’t gain as much understanding of the problem as you would if you wrote the report yourself. So he says you should then advance your understanding of the problem by writing about it.
So he is talking to people like me, who use ChatGPT to answer complex questions. I agree that my understanding improves when I don’t just read ChatGPT’s output but then use the knowledge gained in further writing.
Martin, I have to disagree about what he is saying.
I think “understanding” is a handwave, an evasion. I’ve seen it used that way so many times that I’m quite confident in saying that.
How about doing what I suggested and giving a definition and a description of how we test for it?
Your last sentence says “use the knowledge gained in further writing”. How do we distinguish “understanding” from “knowledge”?
I have lived in Norway for 30 years now, but I am still not fluent in Norwegian. Norwegians use this phrase all the time: “Vær så god.”
I know the literal translation is “Be so good,” which is a polite request for assistance. I also know that is not what Norwegians mean when they say Vær så god. It has multiple meanings in Norwegian. It can mean, “Here you go,” when someone puts food on the table for you. It can mean “After you,” when you and another person try to go through a door at the same time. It can mean “You’re welcome,” when someone says thank you. There are other meanings I’m not recalling at the moment.
I know the literal translation. I know that the literal translation is not what Norwegians mean. And after too many years of observing Norwegians use it in different situations, I know what Norwegians mean when they say it.
Understanding refers to the transition from the incorrect literal meaning to one of the correct Norwegian meanings. When I say I understand what Vær så god means, my brain executes the transition from incorrect to correct again. It happens instantly; I don’t have to think about it. But if I do dwell on it, I see that is what is happening. I am transitioning from incorrect meaning to correct meaning.
MS,
Hmm. I would think that “Vær så god” is context sensitive, and that context can be enumerated. It would come closer to “understanding” if a machine could invent a completely new and novel use of the phrase that could be instantly comprehended by a native speaker and still sound natural. In that sense understanding would be contingent on a creative act of “thinking” based on a deep knowledge of what humans know and how they interact socially.
Or not. Just me spitballing.
Martin, what you describe sounds very much like “know-ledge” to me, and it would seem a trivial effort to have a machine do it. You know what the words mean for each situation. That it may happen instantly without awareness doesn’t change that; it’s obviously true for all the other words we use as well. It’s called proficiency in a language.
If you ask for a definition of “understanding”, even from just the Google search, it is much more complicated. (I suggest doing it.)
What you described… writing some more using what you just learned… isn’t about the typing, but rather that you are continuing your education on the topic. The more you creatively immerse yourself in a subject area, the more likely you are to develop the abilities described in the Google definition of understanding.
My objection to the video is the suggestion that writing or doing long division or whatever is sufficient to achieve the magical state, because in my experience it isn’t even necessary. It’s a basic correlation/causality fallacy that is all too common.
Zebra & Radge, what Curt and I mean by “understanding” is the process of coming to know, or, the process of knowing. Knowledge is what you know.
Know and understand are verbs. Knowledge is a noun. Curt doesn’t mean writing is the only way to understand something. He’s saying that in the new age of AI, reading what AI produces isn’t enough to understand it. Reading and writing use different parts of the brain.
e.g. reading a description of how to hit a 100 mph fastball is not sufficient for you to understand how to hit a 100 mph fastball.
Hi, Martin,
“He’s saying that in the new age of AI, reading what AI produces isn’t enough to understand it.”
I’d say that is also true of reading and writing generally. Reading, depending on how you do it, is more passive than writing. The revisions you make while writing should include clarifying your thoughts and the potential to ditch or change your own opinion. I think that is evident by simple self-inspection. Physically writing also causes you to slow down and massage your thoughts manually into place. Maybe that’s just me.
And don’t get me started on the difference between reading on the Internet and reading a well laid out book, in proper lighting…
I just question what you mean by understanding. Knowing how to hit a ball is one thing. knowing and playing by the rules is a step up from that, being able to rewrite the rules to improve the game comes closer to understanding, IMO.
Just as an aside. I think equating the functioning of neurons only with gateways etc. as an analogue to human intelligence ignores how human thinking may involve intuitions derived from how the brain communicates with the body through hormones and glands etc., and also sensory input (proprioception, vision, and all the rest), through for instance, balance, proportion, harmony– to aesthetics and the experience of a beautiful solution. (Maybe I should rewrite that sentence. I wonder if that would cause me to scrap it…)
Anyway, I suppose all that could be modeled in a computer, but would there be any point in it?
What an LLM is good at in terms of understanding –:it can take some old software library that I had written which was poorly commented and documented and could create a driver for it and figure out its intended use. Once I looked at the LLM response, I realized it was from a paper I had published a while ago. So, it’s essentially like asking an underling to go out and figure something out that you yourself don’t want to rack your brain over.
If that isn’t about using an LLM to aid your own understanding I don’t know what is –as it’s basically pulling your own original understanding back from the brink. I don’t agree with that YouTiube commentary because of my own experiences. It’s a big paradigm shift
.
Here’s an example of a large project that I had used for my personal use but now pulled it together with the aid of LLMs: https://github.com/pukpr/AdaPACE
John Von Neumann said: “You insist that there is something a machine cannot do. If you tell me precisely what it is a machine cannot do, then I can always make a machine which will do just that.”
I contend that understanding is what we cannot program computers (or AIs) to do because we cannot say precisely how it is that we do it. It isn’t simply knowledge. A computer can know whether it knows it knows or not. A honey bee can “know” where a desirable nectar can be found–and transmit that knowledge to the rest of the hive even though it lacks a central nervous system.
In writing, knowledge is knowing that there is a right word for a sentence or sentiment and knowing that right word. Understanding is Mark Twain saying: “The difference between the right word and the almost right word is the difference between the lightning and the lightning bug.”
We can know the history of Germany in the 1930s and see parallels between that era and the US now. Understanding is Mark Twain saying: “History doesn’t repeat, but it rhymes.” It’s telling us something that merely observing the similarities doesn’t.
Knowledge is the skills acquired by Picasso that allowed him to paint Science and Charity at age 15. Understanding is what allowed him to paint Les Demoiselles d’Avignon or Vase with Flowers 10 years later.
Knowledge is understanding the seeming increase of mass of electrons/holes under the influence of some atomic lattices. Understanding is Philip Anderson seeing the similarity between that and the different masses of two otherwise identical leptons (electron and muon) and glimpsing the Higgs mechanism for symmetry breaking and the generation of mass in particle physics.
I cannot say how a human mind reaches understanding. The same 12-20 watts that produce these profound insights also give rise to the babblings of a Donald Trump or a RFK, Jr. (or John Nash who experienced both the profound insights of discovery and the disordered thinking of schizophrenia). It may be an emergent phenomenon. It may be just a matter of time until it emerges from an AI, but it isn’t something we can program, and it is much more than mere knowledge.
Martin,
What you are describing is using the context of a situation to resolve ambiguity, which is what happens when you interpret that Norwegian phrase at various times. The phrase can have different meaning, but your brain quickly determines which one it is from the context of the speaker. The LLMs work the same way — both the training phases and the prompt are used to disambiguate meaning from a prompt you have supplied.
Incidentally, it was back in 2013 that I added the term Context Model to Wikipedia. Those were the days of achieving the pipe dream of resolving ambiguity in the process of automatically generating software and system designs. Needless to say, process has been rapid since that time. .
Martin, your title for the video is “why write?” So when you say…
“reading a description of how to hit a 100 mph fastball is not sufficient for you to understand how to hit a 100 mph fastball.”
…, does that mean that writing a description of how to hit the ball would make you “understand” how to hit it?
This brings us back to my point about the vagueness of the term. How could we actually test your hypothesis? By having people try to hit fastballs? But I doubt very much that baseball players write essays to improve their skill; what they do is get descriptions from coaches and other players of “what works” and try it out.
The suggestion from Radge is much more in line with the the definition of “understanding” that you get with a Google query. It means that the person is capable of creatively applying the knowledge that has been acquired.
This has nothing to do with “the age of AI”. You can read what AI writes, or you can read human-written sources. You yourself said that the AI made it more efficient to do the research. But the fact that you then write something doesn’t mean that you “understand” anything.
You could just be re-arranging other people’s words, which doesn’t mean you’ve done anything creative with the subject matter.
Zebra: …, does that mean that writing a description of how to hit the ball would make you “understand” how to hit it?
MS: We have established that you don’t understand understanding. I don’t think I can contribute any more to your understanding of understanding.
I did write that writing about a subject is NOT the only way to gain understanding of that subject. I suppose you never played the game, so it is understandable that you don’t understand that writing about how to hit a fastball won’t give you the understanding you need to hit a fastball. Ironically, writing about the description you read about how to hit a fastball will help you understand the description of how to hit a fastball.
That was Curt’s point and the point I was trying help you understand. In a nutshell: To gain understanding of an output from AI about a complex ontological, cosmological, philosophical, or physical question (i.e. the subjects Curt interviews scientists about and writes about himself), you have to try/struggle to use the AI output you read. For those subjects, the usual way we users of AI try/struggle to use the output of AI is by writing about what we read.
Zebra: It means that the person is capable of creatively applying the knowledge that has been acquired.
Zxactly. That is what Curt said, but you had the sound turned off, and that is what I tried to say as well. I apologize for confusing you by trying to write about it.
PP: What you are describing is using the context of a situation to resolve ambiguity, which is what happens when you interpret that Norwegian phrase at various times.
I agree. I was trying to explain how I came to understand which meaning was being used. It wasn’t hard, nor did it take long. The process of understanding is sometimes easy and quick.
RL: I contend that understanding is what we cannot program computers (or AIs) to do because we cannot say precisely how it is that we do it. It isn’t simply knowledge.
I agree. I think understanding X happens when you experience awareness that you know X, as well as awareness that you know how to/can use X correctly.
Seeing all the comments just above mine in reply to Martin, I can’t resist bringing up a point I’ve made before about the term “hallucination”.
How are they different from what a person does when they produce a new idea or interpretation in a subject area?
I’ve had a handful of occasions where I did things that were judged “creative” and positive. But I’ve also come up with an equal (at least) amount of really dumb shit failures. Which is why I often say “you can’t make good art or science without making bad art or science”.
The best advice I can give anyone who is trying to achieve something new and hard is:
“You have to let go.”
If you don’t let the AI do that sometimes, you are going to miss out on some good stuff.
Reviews of the book, Picasso’s Guernica: The genesis of a painting by Rudolf Arnheim
https://www.researchgate.net/publication/240493095_Review_of_Picasso's_Guernica_The_genesis_of_a_painting
Also an oldie by Arnheim. Some “out there” musings
Entropy and Art
An Essay on Disorder and Order
https://www.ucpress.edu/books/entropy-and-art/paper
A modern take. Science meets art? Um, you decide. It’s kind of fun though.
https://www.limelight.art/portfolio/entropy
In reducing ambiguity, it was Tim Berners-Lee, the inventor of the WWW, who advocated for a thing called the Semantic Web. This was a formalized version of web pages, which had definitions and schema for much of the content. If we had that in place right now, the amount of energy used in contextualizing everything would be greatly reduced. So the tradeoff is energy use versus organizational maintenance. What won out in the end was training of LLMs, because maintaining semantic correctness was deemed too much work. So we have to deal with hallucinations, while the stuff that is already semantically correct, such as software, we get the benefits in reduced software development costs (or at least barriers),
Martin Smith and Ray Ladbury, first thanks for your interesting comments on AI. However chat gpt and google gemini certainly look like they understand things given the way they answer questions. At least they seem to have a certain level of understanding. If it looks like a duck and quacks like a duck its probably a duck. Perhaps consciousness is just something different that enables us to be aware that we understand.
PP:: This was a formalized version of web pages, which had definitions and schema for much of the content. If we had that in place right now, the amount of energy used in contextualizing everything would be greatly reduced.
MS: We could use AI to reverse engineer the web to be that way now. It would use a lot of energy to convert every page, but then it would be done, and then any document to be published on the web could be submitted to an AI service to create the formalized version. once.
Could do that with Wikipedia, for example Wikidata is a version of semantic web. The issue is that whenever humans get involved, most will not have the discipline to annotate with an ontological classification.
MS: We could use AI to reverse engineer the web to be that way now. It would use a lot of energy to convert every page, but then it would be done, and then any document to be published on the web could be submitted to an AI service to create the formalized version. once.
I don’t know about any of this, but I will say that there’s a lot, LOT of junk on the web! Old or outdated pages that were either never meant to be permanent or became irrelevant with newer information and are now disused, broken links, 404s (etc). plus a lot of just plain garbage (maybe portions of the dark web too?) etc. etc. Just a waste of energy to maintain I would think.
This is not to say it should only be “official” and conformist or uncreative and inhuman at all. But still the web could really use a good cleaning up and paring down
re Radge Havers https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846520 – great point! Of course a human brain is typically trained on sensory/etc. stimuli from multiple channels, internal and external (hungry, blue, circle, fast, over there, loud, even tempo, smooth, cold, ouch!, sweet, sour, sad, etc.) (and the brain learns to interpret some of these inputs by combining other inputs – ie have to reach farther to grab that thing (interactive experiences), etc.), so when we learn words, for at least some of them, we learn their meanings from associations with parts of our experiences. I can imagine an LLM being trained with some audiovisual files (stereoscopic?), but smell?
Martin,
Originally, you produced a very logical and well-structured description of the benefits of using AI. But here you are avoiding the question that I have asked:
“I’ll stop there and see if anyone can offer a definitive test, free of confounding variables, about the output of the LLM v the output of a good writer like Martin.”
So far I haven’t seen one.
I have referenced the Google output on the term “understanding”, which is how I and others understand the term. The way you and Curt are using it is to describe your internal, subjective experience. But it doesn’t matter how hard you work; what matters is whether there is an external objective test for what you claim to achieve.
I think Curt understands marketing and performance very well; “keep on writing” can be translated as “keep on watching my show”.
But there is no reason that a person reading an AI product would not be able, as a result, to achieve a new cognitive state. Sometimes, it’s just a missing piece that’s needed, as in Ray’s example of Philip Anderson.
Z: But there is no reason that a person reading an AI product would not be able, as a result, to achieve a new cognitive state.
MS: Neither Curt nor I disagreed with that in the first place. I don’t know what question you want me to answer that I have not already tried to answer. I can restate what I see as Curt’s point, which I agree with: When you use AI to get an explanation of a difficult subject, after you read the AI explanation, using the knowledge you got from it in your own writing about it is one way to increase your understanding of it.
Zebra, It’s more than just a missing piece. The free-electron model of conductors dates from the first decade or the previous century, and the concept of effective mass for charge carriers was fully developed just after WW II. Countless physicists learned and used these theories. A few of them probably even had the breadth to also be following what was going on in particle physics. Anderson saw a possible mechanism to explain the similar patterns in charge conduction in semiconductor and multiple generations of leptons.
Countless people had observed the patterns of the planets, etc. in the sky, but it took Newton to realize that the falling apple was a limiting form of the same trajectories. It allowed Newton to say that he had “calculated” that the motion of the planets about the Sun was “an elipsis” long before he had actually calculated it–his mouth writing a check that it took his brain 3 years to cash.
Humans can sometimes move past what the data tell us because we see patterns. Now we also see patterns that aren’t there–and that is true for smart people as much as it is for dumb ones. What we need is an AI that fosters our ability to see real patterns and keeps us from running after illusions. I don’t think LLMs are that.
MS: That’s a terrific summary. I also deeply appreciated Data’s link to Bernie talking to Claude.
As to repeating myself, you do a fair amount of that too. The art of listening and reading for full comprehension is also disappearing under the invasion of media and advertising for profit without regard to public costs.
Re solutions, a few suggestions:
Attend No Kings protest (28th). Attend protests against warehousing immigrants. Join the protests against data centers which are deeply polluting. Elect progressive candidates. Stop enabling tech bro groupthink and. idolatry.
Return to progressive taxation. I’m not for a one-time billionaire tax, but for returning at least to the rates under Reagan, which enabled extreme wealth at the expense of our common benefits. Of course, the enlightenment after WWI & WWII was enhanced by we had a maximum tax rate (Eisenhower) of 91% which made acquisition as a goal less appealing. Greed is good wealth and the abandonment of morality has distorted our society beyond recognition and supported extreme inequality. We are the only advanced nation without some form of public health care, but that’s only one item in a disastrous list.
Call lies lies, don’t support bullying, don’t surrender in advance, call out corruption and predatory extraction. Stop maga from firing experts and getting rid of knowledge, wisdom, compassion, etc. etc. Treat women and other races as equal fellow humans.
—
Here’s another revealing misuse of big data: Sucker: My year as a degenerate gambler — https://www.theatlantic.com/magazine/2026/04/online-sports-betting-app-addiction/686061/?gift=n_xMk-Ah7x90dsBbxIEpERM0HRs-8sLtV__lfWjD7I0&utm_source=copy-link&utm_medium=social&utm_campaign=share [free link]
What to make of this?
https://x.com/LeonSimons8/status/2034655055295467904
And why don’t similar records show up in ERA5? I understand that there are differences between how these two datasets are created (different methodologies), which is probably a vast topic on its own, but I don’t really get the reason for such posts, especially given CR most recent data is preliminary for up to 2 weeks. Am I missing something? Granted, these records will fall sooner or later given the recent trends (2027, perhaps?).
On a side note: have there been any updates regarding 2023 anomalies? It’s been almost three years already – have anything new come up regarding its causes?
For decades, the corporate media have under-reported global warming and climate change, while falsely legitimizing the fossil fuel industry’s global warming deniers as “skeptics”.
And now, as the warming accelerates, and the impacts of climate change and weather of mass destruction become increasingly devastating, the corporate media are working harder than ever to BURY the news.
“Since its peak in 2021, global news coverage of climate change has dropped 38 percent … Last year, the first of Trump’s second term, major broadcast networks in the U.S. cut their climate coverage 35 percent compared to the year before … While the New York Times has published an enormous volume of articles about global warming, its coverage has plunged, declining by 66 percent since its peak in October 2021 …”
https://grist.org/language/global-heating-climate-news-drought-chaos/
FYI:
“The US-Israel war on Iran is a disaster for the climate, according to an analysis that finds it is draining the global carbon budget faster than 84 countries combined … the first analysis of the climate cost has found the conflict led to 5m tonnes of greenhouse gas emissions in its first 14 days …
“As of June last year, climate scientists estimated humans could emit greenhouse gases equivalent to 130bn tonnes of CO2 to leave us with a 50% chance of stopping the climate from heating beyond 1.5C. At the present rate … that budget will be exhausted by 2028.”
https://www.theguardian.com/world/2026/mar/21/middle-east-iran-conflict-environment-climate
OH MY! That sounds AWFUL! When did they estimate that the sky would hit the ground because of that disaster for the climate?!?!?!? Let’s look at some facts to gain a more accurate perspective.
“The United States, with the world’s largest commercial air traffic system, accounted for 200 million tons (23 percent) of the 2017 global CO2 total.” That is about 3.85m tonnes PER WEEK every week of every year, or for 14 days (that’s 2 weeks for you folks in Rio Linda) that is 7.7m tonnes. So, every 14 days, the US commercial air traffic system emits 2.7m tonnes MORE THAN THE IRAN WAR numbers that you cite.
On the cirrus clouds produced by contrails: “The effect is so large today that it exceeds the total warming influence of all the CO2 emitted by aircraft since the beginning of powered flight.”
Source – this is a good article that is PRO AGW, anti-GHGs:
https://www.eesi.org/papers/view/fact-sheet-the-growth-in-greenhouse-gas-emissions-from-commercial-aviation#2
That article uses the acronym RPKs. Here’s what that is:
https://www.iata.org/en/publications/newsletters/iata-knowledge-hub/demystifying-key-air-traffic-metrics-understanding-rpks-and-asks/
It is a terrible thing that Iran killed a reported 35,000 or more of their own people – probably with gunpowder and that probably contributes to GHG emissions. It is also horrible that Iran has launched thousands of missiles, drones, and other weapons at their neighbors. Democrats REALLY SHOULD ask them to have a greater concern for the environment, particularly AGW. See how they respond to that.
Instead of the current Iran war, would it have been better if we’d have allowed them to produce nuclear weapons? If they used one (analysts say that they would have), how large of a crater do scientists estimate would be created where Tehran is now located when the free world responded to their terror? FOR EXTRA CREDIT: What would be the equivalent GHG emissions of that nuclear exchange?
KIA: “It is also horrible that Iran has launched thousands of missiles, drones, and other weapons at their neighbors. ”
Yes, but Trump started this war. It should have been obvious Iran could retaliate in unpredictable ways. A lot of this is on Trump for starting things.
KIA “Instead of the current Iran war, would it have been better if we’d have allowed them to produce nuclear weapons?”
But Trump hasn’t STOPPED Iran producing a nuclear weapon. You haven’t found the enriched uranium. You know the enriched uranium Trump told everyone had been absolutely destroyed by his bunker busting bomb months ago? Just another of his lies the gullible MAGA crowd believed.
And what if you do find the enriched uranium, what’s to stop Iran making more? You guys never think anything through. You have gone into this with no plan hoping sheer brute force will prevail. But once you leave Iran, they will be back building a bomb more urgently than ever now, because they know now its the only way they can deter a future attack. Your stupidity has guaranteed that.
I wish you weren’t right, nigel, but I believe you are. And actually, the destruction of the rules-based order by this Maladministration is in itself an incentive to nuclear proliferation–the very thing this Iranian debacle was meant to prevent. But I believe I’m starting to repeat myself on that topic–though it does bear repetition, given the danger of yet more unpredictable national actors with nukes.
As usual, your comment is a litany of falsehoods, nonsense, misrepresentations, irrelevancies, and licking Trump’s boots. And it has absolutely NOTHING to do with the article that I linked to.
Quote from the OP comment: “The US-Israel war on Iran is a disaster for the climate,…….the first analysis of the climate cost has found the conflict led to 5m tonnes of greenhouse gas emissions in its first 14 days…..”
My comment refuted that claim with data and a source link. The data shows that US commercial aviation emits more GHGs per week than the Iran war; so clearly the Iran war is not a disaster for the climate at least from a GHG perspective. If you disagree, please provide data and a source link. Putting it another way, if your numbers are correct, the Iran war is a relatively insignificant source of global GHGs.
Nigel: “Trump started this war.” False. Iran has been at war with the US and Israel on and off for around 47 years. Trump didn’t start it. He’s attempting to end it before they get nukes and use them. Unfortunately in life, some people or nations give us no option other than brute force to correct their behavior. Trump’s team pleaded with the Iranians for weeks to make a deal to avoid this war, but they were not interested.
The main reason there is any opposition to the war is because of irrational hatred of Trump, aka TDS; with similar irrational hatred of Israel by some people. Do not feel bad if you have TDS – it’s an epidemic among leftists. Mental health professionals are being overwhelmed by people suffering from it. I do hope those of you with TDS can find the help you clearly need.
https://washingtonstand.com/commentary/americas-new-mental-health-crisis-trump-derangement-syndrome
https://www.texasmonthly.com/news-politics/trump-politics-anxiety-therapists/
https://www.zerohedge.com/political/trump-derangement-syndrome-real-and-its-driving-therapists-crazy
Nigel: “And what if you do find the enriched uranium, what’s to stop Iran making more?” Answer: The US military and our allies who understand the need to stop them, and the consequences of not stopping them. Life isn’t always nice and cuddly – sometimes it’s brutal; wishing it were different will not change that.
KIA @26 Mar 2026 at 3:58 AM
Nigel: Trump started this war.
KIA: False. Iran has been at war with the US and Israel on and off for around 47 years.
Nigel: I’m not going to comment on Iran and Israel because I don’t know that history very well. But claiming Iran has been at war with the US on and off for around 47 years is false. They have had violent encounters, and proxy battles over the decades but nothing that rises to the level of war. There has never been a formal declaration of war. Iran hasn’t launched an military attack on Americas continent.
You cannot define a war as just any exchange of hostilities. It has to be an attack that is large in scale and directed at large parts of a country. You throw around the term war much too lightly. Its like something one sees on a conspiracy theory website.
KIA: Trump didn’t start it (current attacks on Iran)
Nigel: Trump started the current hostilities just in the same way that Russia attacked the Ukraine. Trump fired the first shots. Iran had not attacked America or American bases. There was no proven threat of imminent danger to America, meaning within the next few weeks.
KIA: He’s attempting to end it before they get nukes and use them. Unfortunately in life, some people or nations give us no option other than brute force to correct their behavior. Trump’s team pleaded with the Iranians for weeks to make a deal to avoid this war, but they were not interested.
Nigel: Bombing Irans military infrastructure doesn’t stop them enriching more uranium and making a bomb.
Nigel: And what if you do find the enriched uranium, what’s to stop Iran making more?
KIA: Answer: The US military and our allies who understand the need to stop them, and the consequences of not stopping them. Life isn’t always nice and cuddly – sometimes it’s brutal; wishing it were different will not change that.
Nigel: How do you stop them? You would have to be attacking Iran every year. Even that may not stop them making or getting a bomb. It will bankrupt Americas economy. The current attack has cost tens billions already, and you are running low on bunker busting bombs and other key weapons. according to experts. Or you would have to blow up every building in the country and kill most of the population to stop them making a nuclear bomb. And that is just inhumane.
You should have stayed with the Obama Plan that was actually getting useful results. Instead you people had “Obama derangement syndrome”” and were just opposed to anything he did on principle.
KIA: You have Trump derangement syndrome…
I don’t think so. I’m not opposed to military attacks by Trump if there’s good cause, like an IMMINENT and PROVEN threat, or a genocide like happened in Bosnia. In fact I have some sympathy for Trump taking out Iran’s leadership given they killed a couple of thousands of civilians, but obviously America hasn’t stopped there. That is where I get concerned especially as its leading to fuel global shortages!
For the record I don’t like Iran’s leadership one little bit, but Trump is pushing the world towards absolute chaos.
List of violent encounters between Iran and America:
Operation Praying Mantis (1988),
The USS Vincennes Incident (1988)
1983 Beirut barracks bombing (241 U.S. Marines killed)
IED attacks on U.S. troops in Iraq (603 Americans killed)
The 1979–1981 Hostage Crisis
The CIA-backed 1953 coup against Prime Minister Mossadegh created long-term hostility.
KIA: He’s attempting to end it before they get nukes and use them.
BPL: Then he should have let Obama’s agreement, which included on-site inspections, stand, instead of automatically breaking it because he hated Obama.
Mr. KIA: “Iran has been at war with the US and Israel on and off for around 47 years.”
47 years? Why stop there. How about 1953, when the US and Britain staged a coup to restore Reza Pahlavi as the Shah, ushering in a 44 year reign of terror that created the Ayatollahs. That last attempt at regime change didn’t work out so well, did it?
And as to all that enriched Uranium…where, exactly is it? Iran successfully moved it before the strike last summer, and now nobody knows. It could be anywhere on Earth–literally. So now the enrichment may not even be occurring in Iran, and the threat could come from anywhere.
There is a reason why we have not seen past Presidents didn’t attempt to invade Iran:
1) Military advisers predicted precisely what we are seeing happen in the Straits of Hormuz.
2) There was no confidence that the regime that replaced the Ayatollahs would be any better. There still isn’t.
3) Iran is a quagmire that makes Vietnam and Iraq look like Spring picnics. Osama bin Laden’s wet dream was to get the US and Iran duking it out, as he knew Iran was the stone on which American military power could be broken.
Don the Con has conned you utterly, but unfortunately Cheetolini started getting high on his own supply and would up conning himself.
Mr KiA – can you tells us, if there is ANYTHING your Glorious Leader could do to make you question him and your admiration toward him?
Obviously not his association with Jefferey Epstein, the head of the largest pedofile ring in the history of the United States. Nor Trumps affair with a porn-star when his wife was pregnant. Nor his bragging that his status as a celebrity permits him grabbing women “by their pussies” with impunity. Nor his calling a female journalist asking the questions he didn’t like – a “piggy”, Nor the over-30,000 lies documented by the WSJ just during his 1st term. Nor the mind boggling corruption with his and his family worth increasing by BILLIONS just in the first year of his 2nd term, with opulent gifts from foreign interests and from domestic multi-billionaires to get the government benefiting them. Nor his subverting the Constitution by abrogating the rights of the Congress, and converting DoJ into his private law firm tasked with prosecuting people he has a personal grudge against him, even if it means firing the FBI specialists on Iranian terrorism weeks before attacking Iran. Nor his fanning the insurrection by the MAGAts who tried to block the constitutional transfer of power, and hang Trumps vice-President for staying with the Constitution.Nor Trumps pardoning every single insurrectionist who violently assaulted the law enforcement. I guess you kissed your “Blue Lives Matter” goodbye. Nor his lying about the American responsibility for hitting
Iranian school with a Tomahawk, and killing 164 people, most of them children, which he tried to blame it on … the Iranians. Nor his pathetic cabinet meetings that has turned into an orgy of flattery, in which everybody tries to outdo others in praising the new emperor clothes.
Since NOTHING above has made any difference to you – I ask you again – IS THERE ANYTHING HE could possibly do to start you questioning your blind unwavering support for Trump?
This ballyhoo just in from the D.C. alibi store: the Carny in Chief, flanked by his fatties, bungles his pitch but still knows how to play the clems.
https://bsky.app/profile/atrupar.com/post/3mhxtz7bedl2r
I made a post in Feb that was gaslit and ridiculed. It’s unusual for reality to catch up so fast to make the many points being made then so damned obvious now about how critical they are in the real world economy the scientists and commenters who trash talk here actually here live in .
Data says 20 Feb 2026 at 2:18 AM
[Response: Oh no! Won’t anyone think of the refineries? Those poor little smol beans – they will be helpless! This is beyond stupid. – gavin]
https://www.realclimate.org/index.php/archives/2026/02/epas-final-ruling-on-co2/#comment-845472
If only you could see the smile on my face now. :-) “He who controls the spice, controls the universe”—said one character in the movie Dune, highlighting the vital importance of ‘spice’ in space travel and the empire’s survival.
The same is equally true to the flow of crude oil on our planet: it is the vital substance feeding, moving, building and now destroying the world. Without it our civilization goes bust, and rather quickly.
Those who have the power to interrupt its flow thus control, if not the Universe, but the fate of the world economy for sure.
And while many commentators fixate on the price of oil and the effect it has on inflation or GDP, they completely miss the point. There is much-much more at stake here than how much you pay at the pump—we are looking at shortages potentially far worse and far more disrupting than what we saw during and after the COVID crisis. Buckle up ‘cause this is going to be a wild ride.
Won’t anyone think of the refineries?
Prices of everything will rise, especially food. This is not just through the price and access to oil, but also a long list of essential industrial raw materials and most importantly, to agricultural fertilisers. 30% of the global nitrogen fertiliser exports, 35% of Urea, 23-30% of Ammonia, 44-50% of Sulphur and 20-40% of processed phosphates are currently blocked, unable to reach the market. Production has also been significantly slowed so even if the Straits were quickly re-opened, there will be serious shortages this year.
Shortages drive price rises, but huge numbers of farmers don’t have the money to pay higher fertiliser prices, so crop yields will suffer. Some may turn to less fertiliser demanding crops which will reduce staple grain crop production further and diminish soil health for years to come. Reducing grain output and rising prices will put additional pressure on livestock farming as animal feed prices rise, leading to stock reductions with corresponding increases in meat and dairy prices globally.
The Middle East and North Africa import 89% of their corn and 77% of their rice. Closed routes, lower crop yields and higher prices increase the risk of famine, with further security and instability risks as described earlier.
The timing of the worst ramifications of this looming, and now unavoidable, global crisis and economic downturn will play out over the coming two years. Shocks tend to follow 12 months after the initial disruption, they did with Covid. This is now highly likely to coincide with an El Niño led jump in global temperatures with intense climate and weather extremes. Worsening floods, storms and droughts will hit a struggling agricultural system at its lowest point, multiplying the damage, risks and security outcomes. Weakened economies struggling with disrupted supply chains, economic recession, inflation and resource shortages, may face severe security issues at home and at their borders. Even stable democracies are likely to see cost-of-living crises escalate to civil unrest and riots. Less democratic regions will risk uprisings similar to the Arab Spring with violent protests and deadly clampdowns. It will be fertile ground for grey zone actions, mis and disinformation campaigns and radicalisation by both the extreme left and right.
What many hoped we would not see for decades may be just around the corner. An El Niño kick, boosted by accelerating warming and superheated oceans, could very well hit our fragile civilisation while its down. On the plus side, when the fires go out, emissions will finally start to drop.
We have spent decades treating ‘National Security’ and ‘Climate Action’ as separate issues. In 2026, we see they are the same threat. You cannot have a secure border on a dying planet, and you cannot have a stable society if the soil is a casualty of war.
I think Data may be the one who is gaslighting. It was an honest interpretation of the prior comment that Data was *asserting* that petroleum-based fuel consumption couldn’t be eliminated without eliminating other uses of petroleum (plastics/etc.), which is not true. (Had he simply expressed a concern, that would be fine; I used to wonder about that myself.) We could utilize already-known technologies to shift the output fractions to continue to supply plastics, lubricants, asphalt, etc. as needed, while electrifying transportation and industry. This would reduce the total petroleum and natural gas consumption substantially, so it would be much easier to rely on a strategic subset of total petroleum supply. But that is a long-term shift, and, …
… That said, flexibility to smoothly manage supply disruptions (of anything) such as Wars of Mass Stupidity (WhiMS) requires systems normally operate under capacity … (and PS the US should perhaps mandate some smaller ship portfolio to make better use of our port capacity…)
Some very good explanations of what’s going on with AI up to now. 3 weeks ago discussion.
Neil deGrasse Tyson, Chuck Nice, & Gary O’Reilly dive into the mechanics of thinking, how AI got its start, and what deep learning really means with cognitive and computer scientist, Nobel Laureate, and one of the architects of AI, Geoffrey Hinton.
https://www.youtube.com/watch?v=l6ZcFa8pybE
And yeah, AI thinks and reasons and it does it just like humans do it.
Hinton explains the fundamental shift from logic-based rule programming to the biological approach: building systems that learn the way a brain does. Learn about the history of computer science that led to the breakthroughs we have today. We break down the structure of artificial neural networks and the meaning of AI buzzwords like “deep learning” really mean. You’ll learn about the layering of data processing and how the first layer of neurons might detect a simple edge, the second a beak, and the third a bird’s head.
Hinton explains the light bulb moment of backpropagation, the mathematical breakthrough using calculus to send force backward to strengthen previous connections. Is this a process we share with neural nets? We discuss whether AI thinks like us and explore why AlphaGo succeeded in beating us at our own game by generating its own data and whether LLMs will hit a ceiling as they run out of human-written text. Can AI reason? What does it mean for something or someone to think?
As the science turns toward the future, we tackle heavy questions regarding the massive energy demands of data centers and whether AI can reinvent solar technology to save itself. Hinton discusses the “Volkswagen Effect,” where a model might strategically underperform to avoid being unplugged. We dive into consciousness to ask if subjective experience is just a byproduct of complex perception and if chatbots already possess it. What are the upsides as well as the downsides? The singularity isn’t imminent yet… but the “yet” is doing a lot of heavy lifting.
Some useful links: Why Net Zero Isn’t Working. ->
https://www.youtube.com/watch?v=QaotZsbMsCs
Marcus Sarofim started his own blog, announced here some weeks ago. Here are his items (most recent first)
https://substack.com/home/post/p-191471987 – Understanding Drivers of American Deaths: Lessons from the 1995 Chicago Heatwave and the 2020–2023 COVID Pandemic in the US
https://substack.com/home/post/p-190409677 – Environmental Injustice: The harm that hides in averages
https://substack.com/home/post/p-188651433 – Through the Looking Glass – The De Minimis Deception: Thoughts after the first week, and why 0.037°C is not “too small to matter”
https://substack.com/home/post/p-188143526 – Through the Looking Glass: Turning Words Upside Down in the Endangerment Finding Rescission
Do not, repeat not, ask/demand that I ‘explain’ them. They are self-explanatory.
Humanity added 850 GW of renewables last year and 600 odd the year before. we presently use 28,000 TWh power and converting al the road transport to electric would be arpund anaother 10,000 TWh so you say it isnt working but if eceryone got on board it would work. However the political system and lobbying favours for the moment wealthy interests and fossil fuels are certainly that.
The vast majority of humanity live in sunnier (sunbelt areas) and hence Wind, Solar and BESS can do it as Mike Mann has stated many times.
It might not be working atm but it can.
Basicslly “net zero” isn’t working
1) Because as long as fossil fuels – because they still are cheaper (albeit often because they are subsidized, especially for air traffic and military uses) are the only alternative for heavy machinery etc. Except for drones, all military equipment is driven by fossil fuels, as is almost all transportation, production/building etc., will be preferred. More than eighty pct. of global energy consumption is still fossil fuels, mainly oil, with coal coming second.
2) Because the fossil industrial owners have lots and lots of capital bound in new and relatively new fossil fuel machinery. If this machinery shall be replaced by machinery driven by renewables, we need a lot of political/state intervention. The owners want better profitability if they shall invest more in renewables. By now there are few price and other incentives for them to do so. As long as fossil fuels are cheaper, they’ll be preferred.
What kind of public/political intervention is needed? First and foremost carbon fee and dividend. You put a fee on all fossil fuels., at the source and/or where it’t being imported. This fee should be rising every year. This will mean that all production and service using fossil fuels will go up in price, while anything else will be chaper, relatively seen. To make this socially just, we need the *dividend*, ie. redistribution of the fee by an equal amount to each and every citizen. Thus for those using mainly renewables, this will mean a net economic surplus: they’ll gain more inn dividend than they pay in fee (according to economic research, this applies to the seventy percent of the workforce with lower incomes. Everyone gain more, the lower their income is, the richest, who pollute most, will pay the most, because they produce most of the CO2 emissions, more the richer they are. Lots of statistics from Oxfam etc. shows this.
As long as we don’t make fossil fuels more expensive than renewables, the renewable energy sources will only function as a supplement. The absolute amount of fossil fuel used won’t be reduced, until we make it more expensive. And this we must do, because the fossil fuels in fact are much more expensive than their market prices make it seem. This is so, because the main part of the price paid for using fossil fuels isn’t paid until long after they were consumed, in fact we’re all only beginning to pay the full price now, when we are being hit by the concequences of accelerating AGW like extreme weather and collapsing ecosystems.
Ocean Stratification is Boosting Global Warming as ocean layers decouple a record-breaking 2025 heat pulse prepares to collide with a 2026 El Niño, supercharging global weather and accelerating non-linear warming.
Recent observations however show temperatures no longer following a steady linear path, but accelerating. 1 Foster, G., & Rahmstorf, S. (2026). Global warming has accelerated significantly. Geophysical Research Letters, 53, e2025GL118804. https://doi.org/10.1029/ 2025GL118804
Satellite observations of the Earth’s Energy Imbalance show continued increases in the amount of energy being absorbed by the planet.2 Earth’s Energy Imbalance: Quarter century update Tom Harris 25 Feb
The three year temperature average for 2023-2025 was over 1.5ºC and 2025 itself was more than 0.25ºC warmer than the 5 years leading up to the 2023 El Niño.
There is however another worrying trend emerging from the data, that of non-linear step changes or jumps in global temperature. The first of these was seen in 2015-2016, the next is unfolding today following the 2023-2024 jump and another one, is potentially approaching in 2027. The jumps we are seeing are linked to El Niño events in the Pacific Ocean.
In 2016, following the warming spike from a very strong El Niño event, temperatures didn’t return to the baseline trend during the subsequent La Niña.3 Li, X., Li, Q., Wild, M. et al. An intensification of surface Earth’s energy imbalance since the late 20th century. Commun Earth Environ 5, 644 (2024). https://doi.org/10.1038/s43247-024-01802-z
The same thing happened following the 2023 El Niño. As expected, 2024 was the hottest year on record, but 2025 did not drop to the prior trend line, despite a relatively weak La Niña phase that has run to this day. The three year temperature average for 2023-2025 was over 1.5ºC7 and 2025 itself was more than 0.25ºC warmer than the 5 years leading up to the 2023 El Niño. Copernicus: 2025 was the third hottest year on record. Copernicus, Climate Change Service. 2025. https://climate.copernicus.eu/copernicus-2025-was-third-hottest-year-record
Despite humanity’s interest in surface air temperature, the oceans are the real story of global warming. About 90% of the trapped energy and therefore the warming, is in the oceans of the world. Water has a very high specific heat capacity, meaning it takes approximately 3,400 times more energy to heat the same volume of water by 1ºC than air. The oceans therefore act as a huge heat sink or buffer, regulating the climate system.
The core of the crisis lies in something called Ocean Stratification.11 Cheng, L., Li, G., Long, SM. et al. Ocean stratification in a warming climate. Nat Rev Earth Environ 6, 637–655 (2025). https://doi.org/10.1038/s43017-025-00715-5
In these areas horizontal density gradients are bent upward to the surface. It’s like a slide where denser water can slide down and spread out. Water mass formation is strongly influenced by surface winds and density stratification.12 Li, Z., England, M.H. & Groeskamp, S. Recent acceleration in global ocean heat accumulation by mode and intermediate waters. Nat Commun 14, 6888 (2023). https://doi.org/10.1038/s41467-023-42468-z
New observations indicate that this process could be jeopardised.13 Multiple studies indicate an emerging decline in winter mode water formation in both the North Pacific and North Atlantic.14151617 If mode water mass formation rates do decline or even break down in some areas, it would speed up heat accumulation within the top 300m, amplifying surface warming. This is one example of how changes in vertical mixing and upper ocean heat content can drive the non-linear spread of Marine Heat Waves over the extratropical oceans.
This leads to dangerous non-linear feedback loops:
Heat Trapping: Stratification is increasing fastest in the upper layers of the oceans which shifts the ratio of ocean heat uptake to shallower depth’s over large areas.
The Cloud Gap: As the surface warms, Sea Surface Temperature (SST) and atmospheric circulation patterns change leading to contracted tropical convection,343536 supporting poleward displacements of narrowing storm tracks, and shifting SST patterns,37 regulating cloud cover over large regions. The net effect is a decline in cloud cover over the oceans.
Wind-stratification feedback: As the lid strengthens, the winds need more energy to physically mix the upper ocean through wave dynamics. As the oceans stratify, basin wide warming signals appear together with weakening surface winds.
Looking at the Ocean Heat Content (OHC) observations at different depths it is clear that the shallower waters in blue are heating far faster than the deeper waters. The term Marine Heat Waves (MHW) has only been defined in the last 10 years. Since then their prevalence has increased rapidly
One such example from 2023 was the MHW in the North Atlantic which persisted for over a year. A first of its kind for intensity, scale, and persistence.56 During its existence, cloud cover was reduced, winds declined and stratification increased. These factors acted as feedbacks strengthening and perpetuating the MHW.
The energy trapped by stratification eventually has to go somewhere. As ever more heat is accumulating within the top 300m, more of it escapes into the atmosphere as water vapour, turning the air into a high-octane fuel for storms.
The 2024 Valencia flooding provided a grim “proof of concept.” With Mediterranean sea surface temperatures at record highs, the storm didn’t just follow standard physics; it exhibited Super-Clausius-Clapeyron scaling. Instead of the expected 7% increase in rain per degree of warming, rainfall intensity jumped by 20%. The result was a year’s worth of rain (771.8mm or 32 inches) in just 16 hours.
2026/2027: The next jump in global temperatures
Despite 2025 being a neutral to weak La Niña year, ocean heat content set new records by a substantial margin. Such a jump should have only been possible during an extreme La Niña since colder equatorial surface water accelerates ocean heat uptake while the system loses less energy to space. In the past ocean heat uptake to greater depths slowed down surface warming. This process appears to be stalling. The ocean surface layers now absorb so much energy that is increasingly being trapped within the upper 300m, accelerating surface warming and through its connection with the atmosphere, air warming.
When the El Niño begins to return, part of the stored energy will start to be released into the atmosphere. Potentially in late 2026, with peak values over the winter months. We are likely to see global temperatures jump to +1.6°C or even +1.7°C during 2027.
A deadly principle seems to be emerging. With each tenth of a degree of warming, extreme events increase disproportionally. Persistence of each jump in temperature with no La Niña cooling phases create a very worrying prognosis. We have not experienced these temperatures before and can only guess at the weather extremes, fire season, floods, droughts and crop disturbances that will now emerge. Whatever does emerge is then unlikely to retreat, its just the next step in the escalation of climate damage.
If an El Niño fails to develop in 2026 then unfortunately that won’t let us off the hook. It would just mean that the oceans will increasingly accumulate even more heat (aided by the neutral to La Nina conditions) to produce an even larger temperature jump in a following year when a strong El Niño does eventually emerge. The longer it takes, the stronger the force and worse the impacts. Until then stratification and MHWs will likely further intensify.
It happened during 2014-2016, then in 2023/24, so it will happen again…
A warning seems warranted. We are in the process of triggering an ocean-atmosphere feedback that may well never have happened before in Earth’s history. This is because it depends on the warming rate rather than the pure temperature.
The upcoming 2026/2027 El Niño will not simply be a repeat of the past; it will be a global ventilation event, dumping years of “trapped” energy into a fragile atmosphere. We must monitor the ocean’s “pulse” with the same urgency we monitor the air. The lid is popping and the climate no longer resets, it only moves forward.
Ocean stratification is emerging as a vitally important aspect of the changing climate with major global implications. It is driven by a complex network of feedbacks but has the potential to disrupt the planet’s energy flows in hugely significant ways. Over half of the energy accumulating as a result of our emissions is being stored in the top 700m of the oceans compared to just 2% in the atmosphere.
misc graphs refs etc https://substack.com/home/post/p-189991236
Dat said:
“Ocean Stratification is Boosting Global Warming as ocean layers decouple a record-breaking 2025 heat pulse prepares to collide with a 2026 El Niño, supercharging global weather and accelerating non-linear warming.
This is garbage information. There has ALWAYS been stratification in the ocean via the primary thermocline. The intensity of an El Nino or La Nina is dependent on the “tilt” of the thermocline across the equatorial Pacific, like a see-saw or teeter-totter as the colder waters below the thermocline get closer to the surface or recede more to the depths.
The only mystery is to what provokes the motion. For a playground see-saw, it’s easy to understand as it depends on which side a kid decides to jump on the see-saw.
For the ocean, the explanation is less facile than that. Interesting what the Gemini LMM says:
https://geoenergymath.com/2026/03/24/the-mystery-of-el-nino/
Obviously what complicates the whole thing a lot is the connections between the Asian monsoon system, the ENSO, the oscillation in the indian ocean, the monsoon systems in southern and western Asia and eastern Africa, and further what happens with the circulation around Antarctica, especially now when the sea ice around Antarctica is rapidly disppearing, thereby strengthening the ablation of the ice shelves from underneath resulting from more warm seawater reaching the Antarctic seas from the north.
https://www.theguardian.com/world/2026/jan/17/climate-antarctica-ice-sheets-glaciers-melting-research-affect-sea-levels https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2023JC020506 https://swais2c.aq/media/record-breaking-sediment-core-provides-unprecedented-evidence-of-west-antarctic-ice-sheet-retreat .
Unfortunately and contrary to Karl R. Popper’s famous requirement for good research: falsifiable hypotheses, the tendency in climate science in the years following the Paris “agreement” (which was never a real agreement on anything, just vague symbolic and hollow declarations which everyone was free to interpret and fullfill just as it pleased them… and mainly it pleased most to interpret this as just another welcomed excuse to stay relaxed and not to do very much) has been to replace the this and the precautionary principle with “optimism” or what Michael E. Mann calls “lukewarming” or frankly: opportunistic adaptation to the fossil fuel dogmatism intrinsic to the mainstream economic “theory” of the neoclassical-neoliberalist school, which has lately morphed into the more openly totalitarian/fascist tending and by now leading oligarchic ideology: trumpism-putinism-bin-salmanism-muskism-thielism etc. (Some would even call it epsteinism…) In short: the battlecry “Drill, baby drill!” which originally came from the anti-darwinian and anti-scientific prophet Sarah Palin.
Especially in a country like the US, where scientific research is highly dependent on private economic interests, their financing and not the least: fossil capital’s (lack of) goodwill, it’s obvious that scientific research whichs results are interpreted by strong oligarchic interests as running contrary to them and their aims, very easily risk being censored and suppressed, as it obviously is now being by the trumpian regime. Add to this the strong US current of christian-sectarian religious extremism denying every well-established elementary and vital scientific fact about the geological and biological history of the Earth, and you have indeed a very dangerous mixture at hand, given the dire situation regarding global climate and ecosystems now under immense and growing pressure from the enormously expanded human species, far beyond the carrying capacity of our planetary home.
The result seems to me to be a timid reaction from many if not most leading climate scientists, resulting in not very bold hypotheses (if hypotheses at all) and a kind of “Paris-agreement-dogmatism”. Many seem to be mostly engaged in trying to downplay the fast growing threats from global heating (“warming” has a sound of pleasant summer that has no connection whatsoever to the harsh reality of roaming wildfires, heat domes, extreme droughts, extreme downpours, mudflows etc. that we are now experiencing everywhere and every year. Fx. where I live in central-eastern Norway, there has been four cases of widespread 50-100-year flooding during the last fifteen years. For most people living through such events, the theoretical-statistical finesses around the question: have we passed the 1,5 degree “goal” (an absurd name…) or not? seem just completely irrelevant, and surely they are. Such discussions do not help the public to understand the urgency and relevance of scientific research, that’s for sure. And I guarantee you: you and we all need that understanding now more than ever. Scientists should connect with common people and their problems, not with the oligarchs, not with the media high-class chatter, rumours etc. If nothing else, the Epstein case shows us where that leads. Why was he putting up his honey traps? We know a great deal about on behalf of whom, at least if we want to, but not surprisingly that isn’t a theme in mainstream media, except for some excellent examples https://m.youtube.com/watch?v=RHn66-CMMxg&pp=0gcJCZoBo7VqN5tD . How come?
James Hansen and some others lately seem to have been kind of isolated as “enfant terrible”s in a tone that I frankly find somewhat both childish and anti-scientific. Too much consensus isn’t good for science and free thinking, and it soon degenerates into groupthink. Slander should have no place in science, but unfortunately it surely has, since we are among human beings. How this works has been elaborated upon in great and elucidating detail by the excellent research of Naomi Oreskes. The sociology of science as developed by fx. Pierre Bourdieu also rises some warning signs that have to be taken very seroius in these times. People who want historical insight into how bad this can get, should read the german novel “Mephisto” by Klaus Mann (1936) https://en.wikipedia.org/wiki/Mephisto_(novel) and the german satirical drama/tragic comedy written in 1961 by swiss writer Friedrich Dürrenmatt “The Physicists” https://en.wikipedia.org/wiki/The_Physicists . Especially with trumpism and A”I” in mind, these themes are more important than ever.
Finally an explanation for the unusually rapid, early, extreme jump in global temperatures in 2023: “Physical understanding of the extreme global temperature jump in 2023, J. Mex, C. Cassou, A. Jézéquel, S. Bony & C. Deser, Communications Earth & Environment , Article number: (2026) Published: 19 March 2026.”
Abstract
Global surface air temperature reached unprecedented heights in early boreal fall 2023, surpassing the previous record for year-to-year temperature increase by a significant margin. We attribute most of this temperature jump to the onset and maturing stages of the 2023 El Niño, with some contributions from the North Atlantic. Using a process-based analysis from multiple observational datasets, we show that the uniqueness of the 2023 event arose from the La Niña-like ocean-atmosphere state on which it developed. This background favoured (1) a steep year-to-year increase of Sea Surface Temperature, particularly in mean atmospheric subsidence regions, reducing low-cloud cover and giving rise to a record-breaking change in the radiative budget; (2) anomalously sustained precipitation over high sea surface temperatures in the Western Pacific, fuelling unusual diabatic heating compared to canonical El Nino events. This altogether led to an exceptionally early increase in tropical tropospheric temperature in boreal fall, ultimately influencing the jump in temperature at the global scale.
https://www.nature.com/articles/s43247-026-03382-6#data-availability
“We attribute most of this temperature jump to the onset and maturing stages of the 2023 El Niño, with some contributions from the North Atlantic.”
This was deemed pretty obvious from those who were keeping track of oceanic indices at the time. The 3 major ocean basins all had peaks in their respective index (ENSO, AMO, IOD). The IOD peaking for the Indian Ocean was the cherry on the top. See this chart
https://imagizer.imageshack.com/img924/4863/BGOZdA.jpg
The AMO also had an interesting double peak which made the temperature spike wider. Basically what happens is that each peak was superposed, creating a hefty spike.
Thanks very much Paul. I wasn’t aware of the issues you raise, and I should really have thought of such possibilities. Certainly looks like a reasonable explanation for the 2023 temperature characteristics.
The month ends and a new is begun. The daily ERA5 re-analysis numbers at Climate Pulse shows the global SAT started March on a cool wobble and ended on a warm one. The global March average +0.53ºC sits on the +0.3ºC/dec trend line, as have more-or-less the last three months (Dec +0.48ºC, Jan +0.50ºC, Feb +0.54ºC).
The numbers from the Uni of Maine Climate Reanalyser show March with a warmer SH and a cooler NH.
The instrument record being SAT/SST may show a warmer average. The ERA5 60N-60S-SST numbers at Climate Pulse show a warming Dec-to-date. This SST warming Dec-to-date has now reversed all the cooling Jan-Nov last year. Such a warming should be expected about now with an El Niño in the offing.
The El Niño forecasts show it as being a bit stronger in the March forecasts (March NINO3.4 forecast hits +1.5ºC by year’s-end in the dynamic models when the Feb forecast saw it at +1.0ºC). Latest NINO3.4 measurement show the March average creeping above zero.
So will this new El Niño come with “bananas!!” attached? The The Banana!!! Watch continues.
To Tomas,
find another installment in the series of non-radiative forcings
https://essopenarchive.org/doi/abs/10.22541/essoar.177005980.01532748/v1
“Reduced Evapotranspiration and Associated Warming Increase Moisture Convergence but Decrease Precipitation Over Land” [preprint]
Benjamin G. Buchovecky, F Hugo Lambert, Claire M Zarakas, Marysa M Laguë, Charlie D Koven, Inez Y Fung, David B Bonan, and Abigail L S Swann
This time another counter-intuitive result using perturbed parameter ensembles. The PPE is interpreted like a non-radiative forcing experiment that generates feedbacks and associated changes to TOA net radiation. In the previous installment Lague used CESM to teach that suppressing ET generates feedbacks that increase atmospheric water vapor. This time Buchovecky and co. teach that suppressing ET results in decreasing precipitation over land while moisture convergence increases. That is very interesting behavior. They contrast this with CO2 radiative forcing experiment from a CMIP ensemble which shows distinct and opposite effects over land. They also contrast changes to large scale circulation. This time HadCM3 is included in the experiments in addition to CESM. Obviously they expect distinct signatures in observables based on the nature of forcing from these model experiments.
in Re to JCM, 2 Apr 2026 at 2:41 PM,
https://www.realclimate.org/index.php/archives/2026/02/unforced-variations-mar-2026/#comment-846767
Dear JCM,
Many thanks for the reference – an exciting reading for the Good Friday.
I wish you and all climate science & natural ecosystems protection enthusiasts happy Easter!
Greetings
Tomáš
fyi
Increased transparency in accounting conventions could benefit climate policy
Gerard Wedderburn-Bisshop
Published 11 March 2025
https://iopscience.iop.org/article/10.1088/1748-9326/adb7f2
and
https://iopscience.iop.org/article/10.1088/1748-9326/ad7d21#erlad7d21s4
dialogue
https://www.youtube.com/watch?v=27b_hM7npUA
Thanks Data,
Gerard Wedderburn-Bisshop’s campaign to revive genuine environmental conservation through the emission counting framework finds strong basis in the HadCM3 experiments. When land parameters are perturbed through to 2100 under the A1B SRES scenario, atmospheric concentrations span roughly 590–927 ppm given the same fossil emission. That range is clearly significant and a similar magnitude to the fossil signal itself. The major trace gas implications sit alongside the well-established, direct and indirect realclimate impacts of land degradation in driving unnatural hydrological and temperature extremes. However, I don’t necessarily endorse his vintage of fencing off large tracts, since soil stewardship and land in production are not mutually exclusive or somehow in opposition. Conservation stewardship and agricultural production are, in practice, borne from the same values and implemented by the same people. Land-grabbing and fencing in the name of trans national emission targets dispossesses local communities, ignores the specific geographies of environment and change, and reproduces a kind of well-intentioned but ultimately reductive, even neocolonial, logic.. I suspect he understands all these things, and it’s hard to fault him for the pressure of working within a framework that reduces environment into a form that disregards place. Recasting nature as a fungible commodity in the currency of international carbon markets. Worse still, it reflects a form of scientism where only what can be abstracted into code is considered valid or real. Regardless, I completely understand the contemporary pressures that shape his approach and I consider him to be completely aligned in the goal of realclimate stabilization.
Thank you JCM for your nuanced well balanced insights especially the positive values framing of what Gerard is attempting and constricting frames he trapped with in.
An analogy might be the dominant framing of neoliberal economics and what is valued as a supreme yardstick … which has successfully side-lined every single holistic human based alternative to political economy worldwide. All voices are drowned out in the dominating noise of mass belief systems of how society is supposed to “work”.
For example is — “atmospheric concentrations span roughly 590–927 ppm given the same fossil emission.” — I doubt this will make it through the dominant filters that decide what’s useful and what isn’t regarding climate science narratives. Unfortunately. It’s what human beings too often do on any subject.
for example your final note — “I completely understand the contemporary pressures that shape his approach and I consider him to be completely aligned in the goal of realclimate stabilization. ”
I suspect Martin Niemöller would agree with the sentiment. I’ve realized that continuing to challenge the current, dominant mindset here is a Sisyphean task, and I can no longer invest my energy into it. 99% of my posts are discarded so I’m stepping back from debating the dominant ideology. It’s a losing battle anyway and I need to to stop for my own well-being.
New study that may be of interest: Collapse of the Atlantic meridional overturning circulation would lead to substantial oceanic carbon release and additional global warming, Da Nian, Matteo Willeit, Nico Wunderling, Andrey Ganopolski & Johan Rockström , Communications Earth & Environment volume 7, Article number: 295 (2026)
Abstract
The potential collapse of the Atlantic Meridional Overturning Circulation could profoundly impact regional and global climates, yet its effects on the carbon cycle and subsequently global temperature remain seriously underexplored. Here we quantify carbon cycle responses across different background global warming levels using a fast Earth system model. We find that Atlantic Meridional Overturning Circulation collapse increases atmospheric carbon dioxide by 47–83 ppm carbon dioxide, leading to around 0.2 °C of additional global warming at higher carbon dioxide background levels after offsetting ocean-dynamics-driven cooling. Despite the modest global warming effect, regional temperature anomalies are pronounced: Arctic temperatures cool by ~ 7 °C (60 °N–90 °N), while Antarctic temperatures warm by ~ 6 °C (60 °S–90 °S). This latter response originates from deep convection triggered in the Southern Ocean, which ventilates deep carbon-rich waters. Such long-term equilibrium responses reveal key physical and carbon-cycle mechanisms and highlight substantial regional climate risks associated with an Atlantic Meridional Overturning Circulation collapse.
https://www.nature.com/articles/s43247-026-03427-w#Sec1
Couple of other interesting studies:
Deadly heat stress conditions are already occurring, Sarah E. Perkins-Kirkpatrick, Catherine H. Gregory, Jennifer K. Vanos, Jane W. Baldwin, Haley Staudmyer, Gisel Guzman-Echavarria & Ollie Jay , Nature Communications volume 17, Article number: 2590 (2026)
https://www.nature.com/articles/s41467-026-70485-1#Sec2
Kutney, G.: Climate denial and the classroom: a review, Geosci. Commun., 8, 81–105, https://doi.org/10.5194/gc-8-81-2025, 2025.
https://gc.copernicus.org/articles/8/81/2025/
The UAH TLT anomalies for March have been posted. Globally the March anomaly (+0.38ºC) is marginally down on Feb (+0.39ºC) with bigger changes hemispherically in differing directions – cooling in the NH (from Feb +0.54ºC to March +0.33ºC) and warming in the SH (from Feb +0.23ºC to March +0.42ºC).
The global anomalies have been on a downward path since they peaked in early 2024. The hemispheres have seen a more stop-go path to their cooling,
Quarterly UAH TLT averages since start of 2025
… … … … … …. Global … … … … …NH … … … … … .. SH
JFM2024 … … +0.85ºC … … .. +0.97ºC … … .. +0.73ºC
AMJ2024 … … +0.80ºC … … .. +0.89ºC … … .. +0.71ºC
JAS2024 … … +0.76ºC … … .. +0.90ºC … … .. +0.63ºC
OND2024 … … +0.67ºC … … .. +0.84ºC … … .. +0.49ºC
JFM2024 … … +0.51ºC … … .. +0.66ºC … … .. +0.36ºC
AMJ2025 … … +0.53ºC … … .. +0.56ºC … … .. +0.49ºC
JAS2025 … … +0.43ºC … … .. +0.48ºC … … .. +0.37ºC
OND2025 … … +0.42ºC … … .. +0.52ºC … … .. +0.32ºC
JFM2026 … … +0.37ºC … … .. +0.46ºC … … .. +0.28ºC
What do you think of the figures in this new post on CO2 and temperature:
https://justdean.substack.com/p/human-caused-climate-change-is-unmistakably
“This post presents two figures that are the culmination of an extended effort to build the strongest possible empirical case for what the paleoclimate record shows about CO₂ and temperature. They draw on five independent regression fits across four independent archives and 66 million years of geological evidence. The argument stands on its own merits.”
I know that Gavin was critical of a climate sensitivity of ~ 8-9C back when he reviewed Snyder 2016 (one of the papers drawn upon here) on these pages in 2016. Has anything changed? It strikes me that if Rovang is right there are big implications – any thoughts
[Response: I’d have to look at more closely, but the point with Snyder and similar estimates that use the glacial/interglacial cycles is that the orbital forcing is a different kind of fish, and one that doesn’t fit easily into the global mean forcing/response paradigm. The longer term changes (as we discussed in relation to Judd et al, 2024) are fine – but subsume ice sheet changes/other GHGs etc. as feedbacks to CO2, which are not quite what is happening now. – gavin]