Fair enough. I suppose the useful bit is that lots of ignorant people have gone nuts ("OMG this is AGI!") without any details. That's just hype. But, yes, to your point, there is also some substantive and interesting speculation from more knowledgeable people.
I've run ollama on my macbook, watched a couple videos on prompt engineering, tried out stable diffusion on my phone. I am even working on a startup that is basically a shiny website plus an OpenAI API wrapper on the backend. What do you mean I am not qualified to speculate on what Q* from OpenAI is and how it is transformational for society!!?? /s
Honestly, he has some really good takes mixed with....really bad ones.
I ended up unfollowing him as the curmudgeony approach combined with some of the far out stuff was just a bit too much for me. Doesn't really improve discourse in the field.
Yann has been a refreshing source of reason and common sense with regards to AI safety, regulation and open-source. I wish we had more people like him and less AI doomer cultists.
People like his takes because he gives an authoritative gloss to what they already believe. But his points are usually lacking in argument or rigor. Anyone that essentially expects the public to trust them when it comes to the outcome of AI/AGI should be view with suspicion.
I mean I don’t think predicting the future is something that typically involves rigor. The outcome is pretty clear: whatever makes a ton of money. Probably a trusted friend in your pocket that sometimes helps you buy stuff. The most negative predictions are silly because they don’t involve making a ton of money for anybody.
The point about money is important, but we should also keep in mind most outcomes will make a ton of money for someone somewhere.
Hell, there's wars killing tens of thousands of people going on right now, and a ton of money is changing hands making a juicy business for whole industries.
On the contrary, he gives good arguments about why open is safer and closed is more dangerous whereas other side gives, imho, convoluted arguments and asks for them to be proven wrong (as opposed to trying to prove themselves right).
His arguments in defense of barreling forward with AI are terrible. They have zero chance to convince someone who doesn't share his intuitions/interests. For example: https://twitter.com/ylecun/status/1718764953534939162
How easily smart people convince themselves of what they want to be true with zero self-awareness makes me much more fearful of what's to come.
>other side gives, imho, convoluted arguments and asks for them to be proven wrong (as opposed to trying to prove themselves right).
The question is what should our default stance be until proven otherwise? I submit it is not to continue building the potentially world-ending technology.
Burden of proof is a tragically misapplied concept. The question is what position should be the default given some uncertainty? This isn't decided in a vacuum. The default should be decided based on the facts and utility estimates of various outcomes where relevant. In the case where the issue is of purely academic interest, the burden of proof is on those making the claim. When the issue has real world consequences, then utility estimates of various outcomes are the overriding concern.
When you see a gun on the table, what do you do? You assume its loaded until proven otherwise. For some reason, those who imagine AI will usher in some tech-utopia not only assume the gun is empty, but that pulling the trigger will bring forth endless prosperity. It's rather insane actually.
The default position should be based on risks and unknowns, I don't think the GP was making any recommendation based on personal feelings.
The fact is we don't know how current ML actually does what it does, we don't know what we'll have next month, and we wouldn't know how to recognize an AI or AGI if we developed one. The risks and unknowns are high, the default position should be to not develop the technology unless and until someone proves without reasonable doubt why we can and should do it, and how we'll do it safely.
As the discussion shows, the choice of null hypotheses is debated. Fortunately, null hypothesis statistical testing is not the only scientific framework for hypothesis testing.
> The default in science is that the side arguing a point has the burden of proving it correct
Right, and Yann is arguing the point that AI and LLMs are not or will not be dangerous. Where's his proof? As the parent posters have said, he has none.
Before the era of quantum physics, the idea of a weapon like fission or fusion weapons would've been inconceivable without resorting to magical thinking.
Your argument, and Yann's, is that AGI, or what you call AGI, is a kind of quasi-intelligent golem that, despite being generally intelligent, doesn't have human-level intelligence. Your claim that it will never be human-equivalent, much less trans-human/ASI, is built into your worldview. It's not a conclusion. It's an assumption on your part.
People like you and Yann can believe that if you want, but you have no evidence, because nobody knows what's required for human-level intelligence. Nobody knows whether some kind of system involving neural nets could develop human-level intelligence or beyond. It could involve different architecture or training methods. There's no assumption by AI doomers that AGI will be achieved by a LLM with more parameters or more or better training data.
What is human-level intelligence is a normative and philosophical question, which means it will not be resolved in some objective manner. There's no requirements for us to know, there's only requirements for us to propose in the form of norms.
The only approach we will be able to claim objectively can produce systems with human-level intelligence is procreation.
This kind of reminds me of the history heavier than air flight and those that thought it wasn't possible along with those that thought it would only be possible by flapping like a bird.
Of course over 100 years later we know it's not only possible but manned flight is far more capable than natural flight in almost every metric.
Agency is simply something that business LLM products don't need. That does not mean it would be hard to add in say a robotic setting; all animals have it.
>despite being generally intelligent, doesn't have human-level intelligence
I think this whole quote is riddled with assumptions in this debate.
What is human level? Is it really a "level" or is human level just a local variant in a space of possible intelligence varieties that maybe could be sorted along one dimension of levels of maybe multiple? Can it be super intelligent without being autonomous at all?
I'm not saying that you're wrong in just pointing out where people ought to have a myriad of different assumptions.
How would you define or recognize a level of intelligence or autonomy that is sufficient to raise concerns? Would you be able to recognize it before its too late?
An early GPT-4 test ended up with GPT successfully solving capchas by tricking a TaskRabbit worker into doing it for them [1]. When asked by the worker if it was a robot, GPT decided to lie to the worker and claim it had a visual impairment that made it difficult to solve the puzzle. That sounds like a level of autonomy and social engineering skills that could be concerning to a reasonable person.
Real intelligence will be knowledge that voting is for suckers and the world is run by a congress where the dominant parties are the illuminati and lizard people. Now if an intelligence can determine and conduct the sacred ritual to join the voting pool then we might onto something. But the true test will be when the intelligence again avoids the suckers game and that the world is actually...
Here you are, trying to "outthink" an AI and speaking as if you understand (both the AI and the world of politics at the very least)! Isn't that silly!
> Show me LLM that has reached a level of automony and self sufficiency
Autonomy and self-sufficiency are not the only ways a system can be dangerous.
But even if this claim were true, ChaosGPT proves that some humans will almost immediately set about using such a non-autonomous tool to create a dangerous autonomous agent. This is my problem with LeCunn, nearly all of his points are trivially refuted by real world observations, yet he keeps repeating them as if they simply must be true.
> Also, even as tools they have fundamental limitations which stem from their autoregressive nature
That's yet another speculative point that LeCunn constantly asserts. Scaling laws have not shown any indication of even approaching a limit.
I've read a fair bit of LeCunn and frankly have been very unimpressed. He is, as far as I can tell, the only big AI name who doesn't think that if we somehow managed to make an actual superintelligent AI it wouldn't be dangerous.
And his reasoning for it is absurd, a combination of evolutionary psychology (not the most scientifically rigorous field to begin with and not even his field) and the belief that no one would choose to make an unsafe AI system, while simultaneously arguing for open sourcing everything so that anyone on the planet can do so if they want.
His reasoning is that good guy’s AI will counteract bad guy’s ones or a an AI that has developed his own objective against humanity if that happens, which is I think a reasonable argument.
He effectively is arguing that the "good guys" will develop a superhero AI that will protect us simply because it feels that is the right thing to do. I'm not sure how any logical basis can be used to back that up, or where any meaningful example of similar behavior that didn't ultimately lose to the "bad guys" can be found in human history.
In the 20th century this same thinking lead to MADD and massive stocks of nuclear weapons that still present an existential threat to humanity. Depending on the scaling potential of intelligence this just adds further risks at this level, not less or balanced risks.
At the same time, it seems like some antidote is needed to the breathless, quasi-mystical hype that cryptic OpenAI claims seem designed to stoke. Demanding precise and substantive criticisms of something about which almost no technical details have been provided seems like an unfair bar.
This was a view that was initially taken by the government for encryption as well. But everyone can agree open sourcing the algorithms and libraries has been the best move.
It is the same with AI/AGI. Anything closed source and having regulatory oversight is useless, decreases innovation, increases bureaucracy and will only serve those who wish to build a “moat” to further their hold on the technology.
> Anything closed source and having regulatory oversight is useless
Source isn't even the problem, unless you're a billionaire, you can't afford to train the model.
This isn't the wheel, the printing press, the PC, or the public internet, which created opportunities for everyone.
This is pay-to-play that is only affordable to the nation-state the mega-corporation, and the latter might let you play around as a digital sharecropper on their platform, until they cut you off, because they can make more money by having first-party ownership of whatever you built.
The model is the instance because the instance is what is being used at inference time. The training simply produces some set of weights.
Open source would mean someone could see and run the model code locally/independently to create an instance. For an LLM, this is much less insurmountable of an issue in terms of resources needed.
We should probably be calling for an open model or open data approach if that's what we think is best.
Calling it open source leaves plenty of gray area in the definition.
For example, I run a few instances of open source software though my instances' databases are private.
We are still answering questions on LLM/AI scaling. Ya, you might have your cottage industry AI giving out answers on the trickle of information you feed it, but will that even be comparable to one being fed megawatts of power with terabytes of data per second flowing into its databanks?
Sure, open source worked well for encryption. It wouldn't work well for bioweapons or nuclear weapons technology. The question is which category does AGI best fit? We shouldn't elide analysis for obtuse comparisons to the past.
But the post links to Twitter (X.) That's the wrong place for argument or rigor. But do you really need that? Seems to me like his post is essentially saying "Pay attention to all this stuff which people have been working on for a long time." He's just casting a vote. The argument and rigor is already there for further reading. That's interesting for me, because I don't know much about the topic. If you already know all this stuff, then maybe it's not interesting.
> I wish we had more people like him and less AI doomer cultists.
There are many sources like him, often with even more detail and references. The problem is that these people are not famous and thus do not have as many followers and are less likely to make it into your feeds. Lots of grad students are in this region as they write to help increase their name and visibility due to the need to market one's self. Don't be afraid to look at those who are not from big name schools. Instead look for those that are willing to mention details and nuance. But this may be hard to accurately determine being on the outside, but that's the thing hype people take advantage of. I don't think they're malicious, but if you're the "smartest person in the room" it's easy to think you're the smartest person in many rooms and have a high confidence in bullshit. For example, we see such comments on places like Reddit and HN. I'm not sure if there's a good way to realistically filter out non-experts from experts (I don't think we should require credential checking). But experts are probably usually more boring as is reality to fantasy.
Yann LeCun is a clown lol. The Chief AI Scientist at _Facebook_ of all companies expects me to just trust him when he says "it'll be safe guys, we'll figure it out, we wouldn't release anything that was harmful :)"
It's hard to call anything that comes from LeCun "news": any time you hear of a phenomenon in the ML/AI space, you know pretty much exactly the sentiment he's going to express. His entire brand is "doomers are wrong, you can trust me, I am AI daddy".
And one can be unfalsifiably successful by using mass media to proclaim safety at every instant. The irony is that if things do fail these kinds of pronouncements can no longer be made using tools like Twitter, etc. on which influencers like LeCun build their brands, since their existence and utility depends on the stability of society.
That's a reductive argument, and won't work with me. What we have is tantamount to an arms race, and trying to suppress another country's development of tech that could be used against them is a thing. We have already restricted access to our microchips to China specifically and Russia through broad sanctions. https://www.nytimes.com/2023/10/17/business/economy/ai-chips... Russia isn't poised for the current crop of AI tech anyway.
Sanctions slow down advancement, but they don't stop it. Someone, somewhere is going to say "fuck it" and build it regardless.
What they can do is ensure that the sanctioned entities are always at least a generation behind. That's not especially helpful in the context of "anti-doomers will be wrong exactly once" unfortunately.
Didn't china manage to produce 5nm chips recently? I mean yes, it's not 3nm but this just means a bit of slowdown, but they have both will& resources to go forward. Just look at their advancements in network wide blocking & face recognition, just like how their big players released pretty damn powerful llms
Well, GPT-2 led directly to GPT-3. GPT-3 led to GPT-3.5 and then to GPT-4. GPT-4 might lead to all of us losing our source of income, so they may yet be proven right (economic turmoil can be an x-risk if it's large enough).
Fortunately, humanity is good at adapting. I remember when gpt2 was going to end the world. Many people can now easily identify when a piece of text was generated by gpt4. This is adaption.
I should maybe dig a bit deeper into what he is saying, but every-time I get excited about some development I get discouraged by his views. Perhaps they are realistic, but I prefer to dream sometimes.
I took an AI course in college back around 2015. Just a bit before AlphaGo.
One project was to implement a simple Q-learning action/value system to play simple games, like Pacman.
The crypto-bros-turned-AGI-experts on twitter are spouting the most uninformed, misguided garbage about this whole thing, it's quite amazing to watch.
And I'm not saying that I am smart or an expert about Q* because I took an introductory college course. I'm saying that even I, someone who knows basically nothing beyond the introductory concept can identity that these people have no clue what they are talking about, and yet the have this incredible talent of speaking in such an authoritative and faux-intelligent tone. It's amazing.
My favorites are the tweets that sound like this:
"So, now we know that [insert something totally wrong]. Well, what if extend that further, by [another totally wrong conclusion]. Here's an explanation of how this all works. A thread, 1/N"
followed by a full thread, images included, of drivel.
> these people have no clue what they are talking about, and yet the have this incredible talent of speaking in such an authoritative and faux-intelligent tone.
I miss the good old days with those idiots were mostly stuck to the honeypot that is/was cryptocurrencies.
> The crypto-bros-turned-AGI-experts on twitter are spouting the most uninformed, misguided garbage about this whole thing, it's quite amazing to watch.
After the LK-99 debacle where Twitter "confirmed" its superconductivity and breathless sheisters heralded the dawn of a new age, my new policy is "inverse Twitter": if Twitter thinks something is a big deal, then it's more than than 50% likely that it's unsubstantiated horsepucky. The signal:noise ratio on Twitter has always been uselessly low, but the post-crypto scene has plumbed heretofore unfathomable depths.
How am I, someone with no experience in AI, supposed to make an informed decision? Certainly not by myself. By reading experts on the topic? But which experts? Twitter ones? In the day of misinformation we need some heuristics (or trusted experts), just look what happened during COVID when people with no experience in medicine suddenly decided they can make their own conclusions.
People without the ability to understand the current body of knowledge and make an educated conclusion, caveated with the known and unknown gaps in current knowledge, simply shouldn't be trying to make conclusions.
Making a decision for oneself is totally reasonable, but there should be a high bar as soon as someone tries to push that as a conclusion that others should follow.
The problem during the pandemic response wasn't that individuals tried to understand and decide for themselves, its that those who were supposed experts got out over their skis and were forcing decisions and conclusions that weren't backed up by data. What's wrong with someone reading studies and deciding whether they themselves want to get vaccinated or west masks, especially when we didn't have solid data to show how those decisions might impact others' rights?
If you were selective on Twitter, one of the best guys was saying it's all over about 24 hours after the hype started and way before the general press got there. The trick is to try to filter the experts from the fools.
I'm confused. I thought people were worried about the danger of some AI breakthrough? If researchers at OpenAI have developed an LLM more advanced than GPT4 which can also plan is that not potentially a worrying breakthrough?
I think you're missing my point. LeCun is saying, "ignore the deluge of complete nonsense" while also saying something to the effect of, "they've probably just created an AI model which can plan".
I agree that the reporting on Q* has mostly been nonsense, and I suspect those articles were published just for the clicks, but at the same time this tweet makes zero sense because an AI model which can plan is somewhat concerning so perhaps those articles were on to something?
This is the cake lecture right? I think it gets mentioned at the end of that talk but he might not have said it then. He also has the part about predictive video (which way does the thing fall).
The leak about Q* feels like an olive branch to let the former OpenAI board and Ilya save a bit of face, probably part of the terms for Sam coming back plus it distracts from all the drama and puts a positive spin on things.
>"The board didn't handle things well, but they were right to be concerned because OpenAI did have some sort of research breakthrough"
Not coincidence that this leaks after Sam comes back, rather than before when it could have made the board look more justified in their decision. This changes the story from incompetence to "it's a problem only OpenAI has because they are so far ahead and close to AGI". Masterful PR move to leak this and shift the narrative
In my opinion, the fiasco is still a reliable warning sign and is drawing attention to OpenAi/Sama/Microsoft dynamics. It may keep on giving even if things appear to have stabilized now.
Could it have been due to him breaking some kind of NDA? After all that would make sense then that they're so mysterious about it, after all they don't want to release more information.
If the original reasoning of Sam's lack of honest communication with the board was an indication, it would make more sense if that comment tipped the board off to a breakthrough they weren't made aware of.
That's not to say it must be deeper than the usual power plays and politics, but breaking an NDA doesn't seem to fit well fromwhat I've seen.
I don't buy this at all. Just going to leave a previous comment I wrote about people coming up with semi-outlandish reasonings behind why something occurs, when there is nothing wrong with the straightforward assumption (that is, the board was concerned about the direction of AI, which is wholly unsurprising given the board's makeup, the Q* stuff was just another concerning point, and there were obviously personality clashes): https://news.ycombinator.com/item?id=38350377
A conspiracy is an understanding between people to facilitate action toward a purpose. A theory about a conspiracy is when outsiders to the conspiracy try to reason about motives without being privy to the actual motives (if any) of the conspiracy. By their nature, conspiracy theories can be far more wide ranging than the actual motives of any conspiracy will be. This is one of the things that makes conspiracy theories problematic.
Given that we already know that there was a conspiracy (by the now-ousted OpenAI board to remove Sam Altman), and that people here on HN have been theorizing about this for almost a week now, your definition makes everyone who commented about it a "conspiracy theorist".
In actual "language game" usage, it's just a derogatory term meaning "I don't like these unproven claims, even though I can't disprove them".
No it’s actually a derogatory term meaning “excessively complicated theory to support my priors.”
Such as casting clear evidence toward the “safety/commercialization of next gen tech was a point of contention among the board” thesis instead as:
> The leak about Q* feels like an olive branch to let the former OpenAI board and Ilya save a bit of face, probably part of the terms for Sam coming back plus it distracts from all the drama and puts a positive spin on things.
What if there's multiple overlapping but distinct meanings of the term conspiracy theory? Wouldn't that be a conspiracy in itself to diffract the meaning? Or would would the mere suggestion that it were itself be a conspiracy theory hiding a far less sinister reality? Or what if, like a fractal, it's everything everywhere all at once? :) hahaha
> A theory about a conspiracy is when outsiders to the conspiracy try to reason about motives without being privy to the actual motives (if any) of the conspiracy. By their nature, conspiracy theories can be far more wide ranging than the actual motives of any conspiracy will be. This is one of the things that makes conspiracy theories problematic.
This seems like it would apply equally well to all theories.
Motives are a lot more difficult to reality check than are physical properties. Especially if, as ethanbond indicates, the person with the theory doesn't believe what the other people say about their motives.
You're trying to use a definition of "conspiracy theory" that, even if you exclude natural phenomena -- which you're only managing to do by special pleading -- says that a conspiracy consists of one or more people cooperating with "each other" in public.
And the conclusion you draw is fully general. There are more possibilities for why something happened than there are reasons for it. What are we supposed to have learned?
> Especially if, as ethanbond indicates, the person with the theory doesn't believe what the other people say about their motives.
The modal case is that, when someone makes a claim about their own motives, they are lying.
The second most common case after that is that they don't know their own motives and their reported motive is not accurate.
Quite frankly the wild speculation is the part of this entire event that I have found most fascinating. When it comes to the business side of things I am far enough removed to not really have an opinion. I don't know the internal thoughts of any of the players and I know sufficiently little about them that I couldn't guess. So I'm not in a position to tell what should have happened. I'm naturally inclined to believe that everyone was doing what they thought was right.
In contrast to that have been these wild theories that are mutually exclusive claiming to know the beliefs, emotions, and plans of practically anybody involved.
In all I have found it to reveal far more about the speculators than it has about events themselves.
People always, always want random human bullshit to be intentional, rational, or preplanned by people smarter than the pliable masses. So they claim there's a secret grand motive behind every piece of information that gets leaked.
It's to save face.
It's all a cover to hide a big R&D discovery they made but don't want to reveal, so they made it about personal dispute with Sam.
It's a marketing ploy to hype people up about a new OpenAI product.
I haven't heard any real details of what Q* is supposed to be; I assumed it's related to reinforcement learning. I guess it's possible the heuristic in A* was replaced with a neural network value function. I'm not sure how that could lead to a breakthrough in language models though.
The star in A* doesn't have anything to do with the algorithm itself. It just means optimal.
Q star would mean optimal Q. If that is even what the star is referring to. There is still no way to know if the Q is the quality factor, query value, question, something else, or just a placeholder name that doesn't reference its structure at all
There is virtually zero chance that this is what Q* is. Loosely connecting the fact that both are named with a letter and an asterisk is the only basis there is for thinking this.
A conspiracy minded me may see all that saga of the last week as a marketing campaign to generate excitement for the GPT-5-now-with-30%-more-Q release.
Ane even more conspiracy-minded take - from the Russian news naturally - is that it is the Great Battle for the future of humanity between "doomers" (Oh, no! the AI is going to kill us all, we need to stop all the work and control the GPUs like guns) and "Effective Altruists" (We can do all the evil today in order to achieve greater good tomorrow)
>A conspiracy minded me may see all that saga of the last week as a marketing campaign to generate excitement for the GPT-5-now-with-30%-more-Q release.
Close. It’s to generate hype for the upcoming $86 billion share sale.
Does anyone have any references to published works discussing the idea he’s pushing here that all the top labs are trying to replace next token prediction with planning? (Note: I’m not interested in using a token-prediction LLM to solve planning problems; I’m interested in work that tries to replace token sampling with action sampling by formulating the text completion problem as a planning domain.) Or is that not what he’s talking about…?
If you do search for Reinforcement Learning based decoding in language models (or neural machine translation) you'll find hundreds of papers. It's a super common research topic that has been around forever, way before LLMs were a "thing". Just follow the reference from there. Nothing new here, really.
But doing it on the scale of something like GPT-4 is only something OAI can do, and I don't doubt that there's some breakthrough there. But IMO it's likely that the breakthrough isn't algorithmic, but a result of scale, like most of the other OAI breakthroughs.
They have the scale in terms of hardware, but they likely don't have the infrastructure to experiment efficiently on this scale on LLMs specifically. OpenAI has been building LLM-specific infrastructure for years. Why do you think Google Bard sucks so much? It's not because there's some special magic secret that only OAI knows. It's because you need a huge amount of purpose-built infrastructure and trial-and-error iteration from experiments each of which take months. You can't easily catch up to that in a few months. Even if you throw 10x more staff at it, you are not going to make progress faster.
The only way you could catch up faster is by having orders of magnitude more hardware. But due the global shortage on hardware right now, nobody has that, not even Google.
> They have the scale in terms of hardware, but they likely don't have the infrastructure to experiment efficiently on this scale on LLMs specifically
This reads like wishful thinking, in the absence of a real moat. Do you think the hyperscalers who are buiding data centers as quickly as they can get permits do not have the infra scale to match OpenAI? Who do you suppose has been providing ever-increasing revenue to Nvidia these past few years? They didn't start doing ML experimentation when ChatGPT was announced.
I agree that they aren't specifically giving LLMs all of their attention - their infrastructure still has to support existing, profitable products that bring in real revenue. I disagree that they can't experiment to the same scale.
There's so much attention on this field right now (something like ~100 papers hitting arxiv a day). Is the possibility of being blind-sided by a "next generation" technique really plausible?
Isn't it more plausible that someone is just piecing together a couple key public papers in a novel way?
Looks like he is not saying it cz he has inside knowledge, instead he is just saying his usual I-JEPA line. About which, the best way to convince people is to give them something they can evaluate/get their hands dirty. I-JEPA for now still looks hand wavy / incomplete.
Ah! Is Q* a play on the A* pathfinding algorithm, but with a Q presumably for the Query vector of attention mechanisms? Clever, and actually does reveal something about the project.
One thing that makes no sense about Sam Altman hiding the Q* progress from the board is that Ilya Sutskever was part of the board too. As he is the technical lead and Altman isn't technical at all, how would it be a secret from Sutskever?
NB: I know nothing about these people and don't track it, but my consistency check failed on this claim based on the flood of news over the past week.
I had a look at the Collison response and it read strongly like "mate puts in a good word when asked to" for me. An impression that was reinforced by the 'like' that Sam put on the answer.
Also, I wouldn't call it a myth. I would call it 'questions around his technical knowledge'.
The best way to answer those questions would be for direct evidence of his technical ability to be shown, not references from mates (just as any interview for a tech position requires direct evidence).
I agree, I could have stated myself more clearly. Sutskever is a world-class domain expert in AI. Altman surely is intelligent and has some technical skills, but isn't the doing the groundbreaking technical work at OpenAi. My point was that it seems highly unlikely that Altman was canned for knowing technical secrets from the board (about Q* specifically) that Sutskever didn't know even better.
This kind of myth happens to almost every successful person. The more hatred on the person, the less technical they become.
"Elon Musk doesn't know how to code" is my fav. It's as if coding is some sort of mythical and rare skill. A new grad who just learn to code a year ago is able to intern at FAANG, but Elon Musk cannot possibly learn how to code for some reason.
It makes no sense because nobody but Sam and the board knows what actually happened.
My guess is that they arrived to an inflection point. The company was on the cusp of taking an investment which would have valued it at 80 billion. One of the board members mentioned that self destruction would have been in line with their mission of safety for humanity. It seems like at some point you have to ditch the "we're doing this for the benefit of humanity" act and act like a profit hungry machine with a startup CEO going balls out. In this case, "balls out" won. The board was way too late. The AGI escaped with the creation of the for-profit company.
It obviously wouldn't. What makes the most sense is that Altman's ousting was orchestrated by Sutskever but he is too important to let go to another project or he would already be gone.
Would have hardly been a shock that "after an exhaustive search for the next CEO, we have decided the most capable person to replace the interim CEO is low and behold Ilya Sutskever."
The fact even I know Sutskever signed the petition is just another aspect of this Shakespearean drama, Sutskever doth protest too much me thinks.
which is ridiculous washing of his hands. he was on the board as chairman. he could have issued a dissenting statement on Friday, etc. that tweet that late was at best very weird and useless
Eliezer Yudkowsky should be respected in the field of AI. Agree or disagree with his level of concern over AI risk, he has made fairly accurate predictions so far. Last I heard, he was predicting that we could be as little as 10-15 years from the level of AI development that he considers to be an extinctions level risk.
Again, there's plenty of different conclusions that can be made on the risk levels but that doesn't change his expectation on timing of development.
Are you saying statements like "...and then AI takes over the whole galaxy and comes back to eats us!" doesn't sound rational?
He is basically the next generation's Ray Kurzweil but instead of AI granting us biological immortality now AI is going to cause mass extinction.
Instead of uploading consciousness to some digital nirvana we get the slightest variation on The Matrix. "And then it takes over the galaxy and uses the atoms in our bodies for batteries!!"
With the way AI captures the imagination we can't but help take things to irrational extremes.
Eliezer made early predictions that we would solve the protein folding problem when effectively everyone else said we either wouldn't solve it or it was way off. If I remember right he was making those claims around 2005 or so and predicting wed have it solved in something like 30 years. We solved it in 2017.
For more context, Eliezer does himself a disservice focusing so hard on examples of how things could go horribly wrong but I haven't heard any compelling arguments against his core points.
From everything I've seen from him, his concerns revolve around (1) a lack of a bounding function for the development of AI (2) a view that the alignment problem is unsolvable and (3) the risk that we won't recognize an AI was developed until it is sufficiently more intelligent than us to pose a fundamental risk.
If your prediction is "A.I. is similar to magic and can do anything and will become a god" you would of course think it can solve the protein folding problem... I'm not exactly sure that leads to the conclusion his track record is great. (There are many more things A.I. can't do than can do, I would think?)
His prediction was that humans would solve the protein folding problem, it didn't have to do with AI.
I raised it only as an example of a prediction he made that most at the time thought had no chance. My point isn't to claim he's some genius that gets everything right, but there are examples of him making predictions widely viewed at the time of having no chance that turned out to be correct.
Writing him off because people responding strongly to his predictions of AI destroying us all doesn't do anyone any good. His core arguments and predictions seem sound and I haven't yet found anyone to explain how we can solve alignment after the fact, for example
Again, if my prediction is "every scientific advance will happen" then I can claim prescience by claiming every scientific advance is evidence of my great insight. That hardly makes me insightful.
I'm writing him off because he isn't a practitioner. I can make things up about what will or won't happen on my own, as can anyone else.
If I want to believe every wacky concept from science fiction will happen, I don't need a middleman to tell me it will.
Do you have an example of a prediction he made that didn't own out? I'm sure there are some, I wouldn't expect anyone making predictions to bat 1000, but throwing his opinion out even when there are examples of accurate predictions doesn't seem fair.
It does feel risky to write off anyone that isn't a practitioner. There won't be any practitioners working on AI risk, its a theoretical pursuit by nature. The closest a practitioner could get is AI risk management, i.e. attempting to limit risk for AI we're already developing.
I think the problem that GP is getting at is the philosophical problem of apocalypticism or market perma-bears, which can be summed up in this satire of them: "So and so is a great market analyst! He's predicted 30 of the last 2 market crashes!" A broken clock can still be right twice a day, but it's still a broken clock. Predictions about the future which are so broad so as to be hopelessly incorrect should not be excused because part of the prediction surface ended up being correct; rather, the entirety of the prediction surface and accuracy should be taken into account to determine the overall accuracy of the prediction.
Taken in that light, his being correct about his protein folding prediction feels like cold comfort when taken in context of his predictions, largely unsubstantiated, of AGI becoming a malevolent human extinction force at par with or of greater risk than that atomic bomb. It elides the very real conceptual limitations of current implementations of AI, such as it's inability to recognizing bidirectional associations (Tom is John's son means John is Tom's father), protecting against injection risks, "hallucinations" (inability to differentiate between when it is making something up and when it is not) and a variety of other existential limits towards its own power.
These details, far from being inconvenient nuances that do not detract from his overall point, should actually serve to make us seriously skeptical of how well he understands the risks of the entity he is warning against. A more skeptical observer might even say (and I'm one of them) that much like fire and brimstone preachers, he is trying to create an unrealistic fear of a deific entity that cannot be reasoned against so that he can consolidate power by serving a "solution" that protects against it and curry popular favor. I find his approach to be the distasteful and manipulative approach of the common demagogue. The haughty, arrogant, insecure way that he treats the arguments of his critics (he simply blocks them on Twitter or condescends to them) make me even more skeptical of him. It makes me wonder if on some level he knows that his argument is not being made in good faith, and isn't trying to win an argument in good faith.
Please don't cross into name-calling. You can make your substantive points without that, and it's in your interest to, since the first part of a comment like this discredits the rest.
Fair enough. I could've phrased that better such as "I find his proto-Millenarian approach to be discursively vacuous, unnecessarily polemical, and not productive towards the answering the very real questions regarding negative externalities around the increasing rate of AI utilization in global society."
AGI is very possible. ASI (artificial super intelligence) is gonna come pretty quick after AGI. The question is when and who will distribute it widely to everyone.
When we have AGI, we will find it out pretty quickly since whoever gets there first wants the most money and control they can.
LeCun is one of the rare voices of reason in the AI community, especially among the celebrities.
He’s been consistently voicing realistic statements with regards to AGI (he doesn’t believe it’s a thing), AI safety and doomerism, limitations of LLMs, etc.
A year ago he said that language models would never be able to explain why if you move a table, the bottle on top of it would not fall. He definitely had no idea what was coming with LLMs. And now he is arrogant enough to say "we all knew what capabilities LLMs would have". He has no intelectual humility or honesty at all. I at least admit openly that language models blew my mind.
Lost respect for his opinions massively after seeing how he lied about the topic.
Can you cite a reference for that quote? I can’t find it. He has definitely underestimated language model capabilities but he has been mostly right on the “language models can’t plan” stuff. He has been consistent and correct about his criticism about auto regressive model as well.
He’s said it’s easy to generate this kind of questions that trick LLMs because of the lack of physical grounding in models trained solely on text.
And that’s true now as ever. I also heard him say that training multimodal models on text+image/video would mitigate the grounding issue, and that’s proven to be true too.
As the other commenter mentions, lecun had a terrible track record when it came to LLMs. All his predictions at the time of what LLMs cannot do were proven wrong.
His research at meta is in analytical methods of AI so I'm not surprised. LLMs were seen as anathema to him as he could not comprehend the capabilities that would come from simply scaling up a suitable architecture with compute and data.
It's quite remarkable how confidently incorrect he was proven to be. Now he's going for round two on his takes with AI safety. Clearly he didn't learn his lesson.
I didn’t realize Noam Brown was at OpenAI. He has done really exciting work, and if he’s pushing the envelope at OpenAI, that bodes well for AGI. His work with modeling an opponent’s beliefs/mind in self-play is on target.
207 comments
[ 3.3 ms ] story [ 157 ms ] threadI could not imagine the same amount of searches and guesswork had they called it "P*" instead.
It seems like Q inherently has some mystery to it. Maybe in that is the least commonly used letter and it rarely is used separately from "Qu"
I ended up unfollowing him as the curmudgeony approach combined with some of the far out stuff was just a bit too much for me. Doesn't really improve discourse in the field.
Hell, there's wars killing tens of thousands of people going on right now, and a ton of money is changing hands making a juicy business for whole industries.
How easily smart people convince themselves of what they want to be true with zero self-awareness makes me much more fearful of what's to come.
>other side gives, imho, convoluted arguments and asks for them to be proven wrong (as opposed to trying to prove themselves right).
The question is what should our default stance be until proven otherwise? I submit it is not to continue building the potentially world-ending technology.
When you see a gun on the table, what do you do? You assume its loaded until proven otherwise. For some reason, those who imagine AI will usher in some tech-utopia not only assume the gun is empty, but that pulling the trigger will bring forth endless prosperity. It's rather insane actually.
That’s not a rational argument for why we should be concerned — you merely asserted you were.
Why is catering to your feelings the default position?
The fact is we don't know how current ML actually does what it does, we don't know what we'll have next month, and we wouldn't know how to recognize an AI or AGI if we developed one. The risks and unknowns are high, the default position should be to not develop the technology unless and until someone proves without reasonable doubt why we can and should do it, and how we'll do it safely.
Right, and Yann is arguing the point that AI and LLMs are not or will not be dangerous. Where's his proof? As the parent posters have said, he has none.
LLMs are tools in amplifying individual human intelligence, 100% of automony and will come from their user (human).
Also, even as tools they have fundamental limitations which stem from their autoregressive nature
Your argument, and Yann's, is that AGI, or what you call AGI, is a kind of quasi-intelligent golem that, despite being generally intelligent, doesn't have human-level intelligence. Your claim that it will never be human-equivalent, much less trans-human/ASI, is built into your worldview. It's not a conclusion. It's an assumption on your part.
People like you and Yann can believe that if you want, but you have no evidence, because nobody knows what's required for human-level intelligence. Nobody knows whether some kind of system involving neural nets could develop human-level intelligence or beyond. It could involve different architecture or training methods. There's no assumption by AI doomers that AGI will be achieved by a LLM with more parameters or more or better training data.
The only approach we will be able to claim objectively can produce systems with human-level intelligence is procreation.
Of course over 100 years later we know it's not only possible but manned flight is far more capable than natural flight in almost every metric.
https://promptengineering.org/what-are-large-language-model-...
I think this whole quote is riddled with assumptions in this debate.
What is human level? Is it really a "level" or is human level just a local variant in a space of possible intelligence varieties that maybe could be sorted along one dimension of levels of maybe multiple? Can it be super intelligent without being autonomous at all?
I'm not saying that you're wrong in just pointing out where people ought to have a myriad of different assumptions.
So, more Ringo than Lenon?
An early GPT-4 test ended up with GPT successfully solving capchas by tricking a TaskRabbit worker into doing it for them [1]. When asked by the worker if it was a robot, GPT decided to lie to the worker and claim it had a visual impairment that made it difficult to solve the puzzle. That sounds like a level of autonomy and social engineering skills that could be concerning to a reasonable person.
[1] https://www.businessinsider.com/gpt4-openai-chatgpt-taskrabb...
If it asks for the right to vote?
Here you are, trying to "outthink" an AI and speaking as if you understand (both the AI and the world of politics at the very least)! Isn't that silly!
Autonomy and self-sufficiency are not the only ways a system can be dangerous.
But even if this claim were true, ChaosGPT proves that some humans will almost immediately set about using such a non-autonomous tool to create a dangerous autonomous agent. This is my problem with LeCunn, nearly all of his points are trivially refuted by real world observations, yet he keeps repeating them as if they simply must be true.
> Also, even as tools they have fundamental limitations which stem from their autoregressive nature
That's yet another speculative point that LeCunn constantly asserts. Scaling laws have not shown any indication of even approaching a limit.
People dismiss LLMs because they are not embodied, and lack continuous training. That is to come.
I only see his posts on Twitter but haven't been impressed.
It is the same with AI/AGI. Anything closed source and having regulatory oversight is useless, decreases innovation, increases bureaucracy and will only serve those who wish to build a “moat” to further their hold on the technology.
Source isn't even the problem, unless you're a billionaire, you can't afford to train the model.
This isn't the wheel, the printing press, the PC, or the public internet, which created opportunities for everyone.
This is pay-to-play that is only affordable to the nation-state the mega-corporation, and the latter might let you play around as a digital sharecropper on their platform, until they cut you off, because they can make more money by having first-party ownership of whatever you built.
Open source would mean someone could see and run the model code locally/independently to create an instance. For an LLM, this is much less insurmountable of an issue in terms of resources needed.
For example, I run a few instances of open source software though my instances' databases are private.
We are still answering questions on LLM/AI scaling. Ya, you might have your cottage industry AI giving out answers on the trickle of information you feed it, but will that even be comparable to one being fed megawatts of power with terabytes of data per second flowing into its databanks?
There are many sources like him, often with even more detail and references. The problem is that these people are not famous and thus do not have as many followers and are less likely to make it into your feeds. Lots of grad students are in this region as they write to help increase their name and visibility due to the need to market one's self. Don't be afraid to look at those who are not from big name schools. Instead look for those that are willing to mention details and nuance. But this may be hard to accurately determine being on the outside, but that's the thing hype people take advantage of. I don't think they're malicious, but if you're the "smartest person in the room" it's easy to think you're the smartest person in many rooms and have a high confidence in bullshit. For example, we see such comments on places like Reddit and HN. I'm not sure if there's a good way to realistically filter out non-experts from experts (I don't think we should require credential checking). But experts are probably usually more boring as is reality to fantasy.
Your verbiage isn't in the spirit of HN.
If someone can build it someone will. Laws and impending Doom or not. It's probably going to be better if it's us than Russia or China.
What they can do is ensure that the sanctioned entities are always at least a generation behind. That's not especially helpful in the context of "anti-doomers will be wrong exactly once" unfortunately.
LOL and how is that working?
No reason to believe this, it hasn't been true for any other dangerous technologies like biological weapons or nuclear power.
One project was to implement a simple Q-learning action/value system to play simple games, like Pacman.
The crypto-bros-turned-AGI-experts on twitter are spouting the most uninformed, misguided garbage about this whole thing, it's quite amazing to watch.
And I'm not saying that I am smart or an expert about Q* because I took an introductory college course. I'm saying that even I, someone who knows basically nothing beyond the introductory concept can identity that these people have no clue what they are talking about, and yet the have this incredible talent of speaking in such an authoritative and faux-intelligent tone. It's amazing.
My favorites are the tweets that sound like this:
"So, now we know that [insert something totally wrong]. Well, what if extend that further, by [another totally wrong conclusion]. Here's an explanation of how this all works. A thread, 1/N"
followed by a full thread, images included, of drivel.
I miss the good old days with those idiots were mostly stuck to the honeypot that is/was cryptocurrencies.
After the LK-99 debacle where Twitter "confirmed" its superconductivity and breathless sheisters heralded the dawn of a new age, my new policy is "inverse Twitter": if Twitter thinks something is a big deal, then it's more than than 50% likely that it's unsubstantiated horsepucky. The signal:noise ratio on Twitter has always been uselessly low, but the post-crypto scene has plumbed heretofore unfathomable depths.
No clue why people try to come up with these obviously silly heuristics.
Making a decision for oneself is totally reasonable, but there should be a high bar as soon as someone tries to push that as a conclusion that others should follow.
The problem during the pandemic response wasn't that individuals tried to understand and decide for themselves, its that those who were supposed experts got out over their skis and were forcing decisions and conclusions that weren't backed up by data. What's wrong with someone reading studies and deciding whether they themselves want to get vaccinated or west masks, especially when we didn't have solid data to show how those decisions might impact others' rights?
There may be a breakthrough, there may not, but nothing on the topic is convincing or worth reading.
> Reuters could not independently verify the capabilities of Q* claimed by the researchers.
True, they are a reputable news agency, but the parent is also correct, it's not highly credible.
I agree that the reporting on Q* has mostly been nonsense, and I suspect those articles were published just for the clicks, but at the same time this tweet makes zero sense because an AI model which can plan is somewhat concerning so perhaps those articles were on to something?
https://thealgorithmicbridge.substack.com/p/gpt-4-a-viral-ca...
LeCun: “[Note: I've been advocating for deep learning architecture capable of planning since 2016].”
Reply: “My understanding is Schmidhuber already solved that 10 years ago. Just no-one knows it yet.”
From 2015. https://arxiv.org/abs/1511.05440
Edit: I didn't realize you were sarcastic (－‸ლ)
>"The board didn't handle things well, but they were right to be concerned because OpenAI did have some sort of research breakthrough"
Not coincidence that this leaks after Sam comes back, rather than before when it could have made the board look more justified in their decision. This changes the story from incompetence to "it's a problem only OpenAI has because they are so far ahead and close to AGI". Masterful PR move to leak this and shift the narrative
That's not to say it must be deeper than the usual power plays and politics, but breaking an NDA doesn't seem to fit well fromwhat I've seen.
Interesting claim. We know with certainty that there was a conspiracy to move against Sam Altman. What does "conspiracy theory" mean?
In actual "language game" usage, it's just a derogatory term meaning "I don't like these unproven claims, even though I can't disprove them".
Such as casting clear evidence toward the “safety/commercialization of next gen tech was a point of contention among the board” thesis instead as:
> The leak about Q* feels like an olive branch to let the former OpenAI board and Ilya save a bit of face, probably part of the terms for Sam coming back plus it distracts from all the drama and puts a positive spin on things.
This seems like it would apply equally well to all theories.
And the conclusion you draw is fully general. There are more possibilities for why something happened than there are reasons for it. What are we supposed to have learned?
> Especially if, as ethanbond indicates, the person with the theory doesn't believe what the other people say about their motives.
The modal case is that, when someone makes a claim about their own motives, they are lying.
The second most common case after that is that they don't know their own motives and their reported motive is not accurate.
> conspiracy consists of one or more people cooperating with "each other" in public.
Public doesn't have to have anything to do with it.
> What are we supposed to have learned?
Yes, I overgeneralized. In general the thing to learn is to try to reality check as much as possible.
> The modal case is that, when someone makes a claim about their own motives, they are lying.
Granting this point, a razor would attribute base motivations over complex motivations.
Does anyone really question what the parent commenter was referring to with the phrase "conspiracy theory" given the context of this thread?
In contrast to that have been these wild theories that are mutually exclusive claiming to know the beliefs, emotions, and plans of practically anybody involved.
In all I have found it to reveal far more about the speculators than it has about events themselves.
p.s., maybe bomb us a good idea.
It's to save face.
It's all a cover to hide a big R&D discovery they made but don't want to reveal, so they made it about personal dispute with Sam.
It's a marketing ploy to hype people up about a new OpenAI product.
etc etc
Q star would mean optimal Q. If that is even what the star is referring to. There is still no way to know if the Q is the quality factor, query value, question, something else, or just a placeholder name that doesn't reference its structure at all
Ane even more conspiracy-minded take - from the Russian news naturally - is that it is the Great Battle for the future of humanity between "doomers" (Oh, no! the AI is going to kill us all, we need to stop all the work and control the GPUs like guns) and "Effective Altruists" (We can do all the evil today in order to achieve greater good tomorrow)
Close. It’s to generate hype for the upcoming $86 billion share sale.
Anyone have a good source for learning more about this topic?
https://github.com/AGI-Edgerunners/LLM-Planning-Papers
Here's one titled "Reasoning with Language Model is Planning with World Model"
https://arxiv.org/abs/2305.14992
https://arxiv.org/pdf/2305.20050.pdf
It’s different from the link the other user posted.
This one is specifically about the Q* process
But doing it on the scale of something like GPT-4 is only something OAI can do, and I don't doubt that there's some breakthrough there. But IMO it's likely that the breakthrough isn't algorithmic, but a result of scale, like most of the other OAI breakthroughs.
It's shocking to me that people think Google and Meta don't have the scale to operate at this level.
The only way you could catch up faster is by having orders of magnitude more hardware. But due the global shortage on hardware right now, nobody has that, not even Google.
This reads like wishful thinking, in the absence of a real moat. Do you think the hyperscalers who are buiding data centers as quickly as they can get permits do not have the infra scale to match OpenAI? Who do you suppose has been providing ever-increasing revenue to Nvidia these past few years? They didn't start doing ML experimentation when ChatGPT was announced.
I agree that they aren't specifically giving LLMs all of their attention - their infrastructure still has to support existing, profitable products that bring in real revenue. I disagree that they can't experiment to the same scale.
Google has nearly unlimited resources to pursue this shit.
Isn't it more plausible that someone is just piecing together a couple key public papers in a novel way?
NB: I know nothing about these people and don't track it, but my consistency check failed on this claim based on the flood of news over the past week.
Patrick Collison seemed satisfied with Sam's tech chops [1]. If that isn't enough of a vouch, god help the rest of us.
[1] https://www.quora.com/Is-Sam-Altman-highly-technical-Has-he-...
Also, I wouldn't call it a myth. I would call it 'questions around his technical knowledge'.
The best way to answer those questions would be for direct evidence of his technical ability to be shown, not references from mates (just as any interview for a tech position requires direct evidence).
"Elon Musk doesn't know how to code" is my fav. It's as if coding is some sort of mythical and rare skill. A new grad who just learn to code a year ago is able to intern at FAANG, but Elon Musk cannot possibly learn how to code for some reason.
My guess is that they arrived to an inflection point. The company was on the cusp of taking an investment which would have valued it at 80 billion. One of the board members mentioned that self destruction would have been in line with their mission of safety for humanity. It seems like at some point you have to ditch the "we're doing this for the benefit of humanity" act and act like a profit hungry machine with a startup CEO going balls out. In this case, "balls out" won. The board was way too late. The AGI escaped with the creation of the for-profit company.
Would have hardly been a shock that "after an exhaustive search for the next CEO, we have decided the most capable person to replace the interim CEO is low and behold Ilya Sutskever."
The fact even I know Sutskever signed the petition is just another aspect of this Shakespearean drama, Sutskever doth protest too much me thinks.
// nitpick, it's lo (from look) not low
What they've accomplished so far is really impressive. But until I see AGI for real I won't hold my breath.
Just feels to me like some expert/wizard level PR is going on right now.
To the average person chatGPT is literally magic.
Again, there's plenty of different conclusions that can be made on the risk levels but that doesn't change his expectation on timing of development.
He is basically the next generation's Ray Kurzweil but instead of AI granting us biological immortality now AI is going to cause mass extinction.
Instead of uploading consciousness to some digital nirvana we get the slightest variation on The Matrix. "And then it takes over the galaxy and uses the atoms in our bodies for batteries!!"
With the way AI captures the imagination we can't but help take things to irrational extremes.
From everything I've seen from him, his concerns revolve around (1) a lack of a bounding function for the development of AI (2) a view that the alignment problem is unsolvable and (3) the risk that we won't recognize an AI was developed until it is sufficiently more intelligent than us to pose a fundamental risk.
I raised it only as an example of a prediction he made that most at the time thought had no chance. My point isn't to claim he's some genius that gets everything right, but there are examples of him making predictions widely viewed at the time of having no chance that turned out to be correct.
Writing him off because people responding strongly to his predictions of AI destroying us all doesn't do anyone any good. His core arguments and predictions seem sound and I haven't yet found anyone to explain how we can solve alignment after the fact, for example
I'm writing him off because he isn't a practitioner. I can make things up about what will or won't happen on my own, as can anyone else.
If I want to believe every wacky concept from science fiction will happen, I don't need a middleman to tell me it will.
It does feel risky to write off anyone that isn't a practitioner. There won't be any practitioners working on AI risk, its a theoretical pursuit by nature. The closest a practitioner could get is AI risk management, i.e. attempting to limit risk for AI we're already developing.
Taken in that light, his being correct about his protein folding prediction feels like cold comfort when taken in context of his predictions, largely unsubstantiated, of AGI becoming a malevolent human extinction force at par with or of greater risk than that atomic bomb. It elides the very real conceptual limitations of current implementations of AI, such as it's inability to recognizing bidirectional associations (Tom is John's son means John is Tom's father), protecting against injection risks, "hallucinations" (inability to differentiate between when it is making something up and when it is not) and a variety of other existential limits towards its own power.
These details, far from being inconvenient nuances that do not detract from his overall point, should actually serve to make us seriously skeptical of how well he understands the risks of the entity he is warning against. A more skeptical observer might even say (and I'm one of them) that much like fire and brimstone preachers, he is trying to create an unrealistic fear of a deific entity that cannot be reasoned against so that he can consolidate power by serving a "solution" that protects against it and curry popular favor. I find his approach to be the distasteful and manipulative approach of the common demagogue. The haughty, arrogant, insecure way that he treats the arguments of his critics (he simply blocks them on Twitter or condescends to them) make me even more skeptical of him. It makes me wonder if on some level he knows that his argument is not being made in good faith, and isn't trying to win an argument in good faith.
How many of his predictions are even disprovable? Do you have a list? Was that protien fold prediction even disprovable?
An example of a nondisprovable prediction is:
-one day an AI will become self aware.
A disprovable one is:
-Skynet will begin to learn at a geometric rate. It will becomes self-aware at 2:14 a.m. Eastern time, August 29th 1997.
https://news.ycombinator.com/newsguidelines.html
When we have AGI, we will find it out pretty quickly since whoever gets there first wants the most money and control they can.
He’s been consistently voicing realistic statements with regards to AGI (he doesn’t believe it’s a thing), AI safety and doomerism, limitations of LLMs, etc.
Lost respect for his opinions massively after seeing how he lied about the topic.
And that’s true now as ever. I also heard him say that training multimodal models on text+image/video would mitigate the grounding issue, and that’s proven to be true too.
So I’m not sure exactly what your objection is.
His research at meta is in analytical methods of AI so I'm not surprised. LLMs were seen as anathema to him as he could not comprehend the capabilities that would come from simply scaling up a suitable architecture with compute and data.
It's quite remarkable how confidently incorrect he was proven to be. Now he's going for round two on his takes with AI safety. Clearly he didn't learn his lesson.