Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why I'm still bearish on LLMs after Navier-Stokes

dank.systems

Post reply on HN
Print view View original

411–420 of 642 posts

Re: Why I'm still bearish on LLMs after Navier-Stokes

#411

Post by solenoid0937 »

datsci_est_2015 wrote: post #401
sailfast wrote: post #380

Earlier quoted context omitted.

What happens when you ask it to play chess against you if the chess game has an API? Are you measuring chess or multi-tasking skill? Also what harness? If you’re using a general harness of course it’s going to try and give you commentary. I say this not because I’m an LLM shill but because false equivalence is all over the place in the space and maybe it’s a fine heuristic for you but probably not a real outcome when…

Why does a 6 year old not need any of these guardrails? Frontier model’s failure modes are a direct refutation of claims that we’ve reached (or will soon reach) the artificial general intelligence. We may have reached an artificial general intelligence, but there may be more complexity to this than even AI thought leaders are talking / influencing about. Maybe not all AGIs have a path to digital singularity. Maybe ou…

Ask a 6 year old to draw a chess board from scratch every turn and they too will make mistakes.

Re: Why I'm still bearish on LLMs after Navier-Stokes

#412

Post by jnwatson »

credit_guy wrote: post #132
iron_albatross wrote: post #104

Earlier quoted context omitted.

When thinking about these valuations, shouldn’t we try to quantify how much knowledge work becomes obsolete if other knowledge workers are automated? I.e. there are a huge amount of knowledge workers employed in businesses that create tools for other knowledge workers. AI won’t automate their work, those businesses will just cease to exist. And then there’s the second order effect: if all the knowledge workers get au…

I think you are committing the lump of labor fallacy [1]. Lots of jobs will disappear, but others will appear. Lots of things (both intellectual and material) that are produced nowadays by humans will be produced in the near future by AI. But humans will be needed to do new things. Take the Hugging Face incident. Why did it happen? Because the people whose task was to set up a testing framework took shortcuts. Why di…

Yes, think of all the new security engineers we'll need!

Re: Why I'm still bearish on LLMs after Navier-Stokes

#413

Post by bulder »

ricky54 wrote: post #256
carodgers wrote: post #17

This April 2026 paper is a fun and related read. https://arxiv.org/html/2509.24239v4 Researchers asked frontier models to play chess. Have a look at the MAR rates in Table 3. When not explicitly told which moves were legal, no model identified legal moves at a rate better than 80%. Many asked for more illegal moves than legal moves. And even when explicitly told which moves were legal, the models continued to ask for…

If you give the same task to an exceptionally intelligent human, who does not play chess and has only heard about it in passing, then they would be beaten by every child who has looked at the rules for more than 10 minutes. What kind of intelligence is "playing but we don't tell you the rules" supposed to test?

...but that's not what the models are. You can interrogate them on the rules of chess, and they'll (statistically likely) give you a decent breakdown of the rules. Evidently the rules are in their training material, they just fail to apply them in the manner of an intelligent system for some reason or another.

Re: Why I'm still bearish on LLMs after Navier-Stokes

#414

Post by foldr »

thom wrote: post #391
foldr wrote: post #320

Earlier quoted context omitted.

I haven't tried it myself, but people seem to report that the illegal moves surface eventually. It just takes longer: https://news.ycombinator.com/item?id=49720751 Nothing is forcing the LLM to play 'blind'. If it's smart, it should be able to create its own representation of the chess board and update it with every move, just like a human would. Any chess engine that's sensitive to how the moves are formatted is cle…

A human wouldn't do that, they'd look at the board. I'm not disagreeing that to demonstrate clear superhuman ability the LLM should be able to do this, but it plays better than most humans blindfolded, and with fair prompts seems very good otherwise.

That's what a human will do if they already have a physical board to look at. But if someone, say, posed you a chess exam question via FEN notation, or as a sequence of moves in algebraic notation, you'd sketch a visual representation of the board off your own initiative to help you answer the question. There is nothing in principle to stop the LLM creating its own board representations in whatever format enables it to easily keep track of game state and legal and illegal moves. If it fails to do so, that's a sign of its own limited understanding of chess as compared to a human.

The LLM would only be playing 'blindfolded' if you somehow forbade it from making notes (as you effectively do by literally blindfolding a human, given how limited human working memory is). But you are not doing that. The LLM is free to keep track of the game state via whatever means it chooses.

None of this is about superhuman ability. Any human who understands a given chess notation can convert it to a visual representation of a chess board and then use that representation to choose their next move, with their usual level of performance.

Re: Why I'm still bearish on LLMs after Navier-Stokes

#415

Post by moron4hire »

joefourier wrote: post #31
carodgers wrote: post #17

This April 2026 paper is a fun and related read. https://arxiv.org/html/2509.24239v4 Researchers asked frontier models to play chess. Have a look at the MAR rates in Table 3. When not explicitly told which moves were legal, no model identified legal moves at a rate better than 80%. Many asked for more illegal moves than legal moves. And even when explicitly told which moves were legal, the models continued to ask for…

> current frontier models > Gemini 2.5 Pro, O3, Claude Sonnet 3.7 and ChatGPT 4.1 The gap in capabilities between those models which they tested, and actual current frontier ones is enormous. I would not trust that any conclusions they made are applicable.

> The gap in capabilities between those models which they tested, and actual current frontier ones is enormous.

Same story every 4 months and yet still no breakout, winning products. I've been hearing "the AI is good now" and "it 10x's my productivity" for a over a year now. If it were true, why aren't the all-in-AI using companies 10-15 years ahead of their competition yet? Why is it still all buggy, poorly designed junk?

Re: Why I'm still bearish on LLMs after Navier-Stokes

#416

Post by cafebeen »

derpyzza wrote: post #394
yshklarov wrote: post #209

Great article, but the lack of sentence capitalization makes it unnecessarily difficult to read. Apologies if this comment is off-topic, but it really is quite egregious, and since the article was submitted by the author I presume they are open to the feedback.

huh i didn't even notice, that's how i write all my blog posts too. it looks nicer to me and i don't have to bother checking for "proper" capitalization if everything's just lowercase anyway. didn't realize people struggled to read text that way though, maybe i should change my writing style if this is a common pain point

I personally find it slower to read text without capitalization. It helps me visually chunk information from different sentences, in the same way that line breaks separate ideas into different paragraphs. Without it, it’s sort of like a wall of text that I have to read more carefully to parse (although a closer reading perhaps isn’t bad).

Re: Why I'm still bearish on LLMs after Navier-Stokes

#417

Post by cbolton »

MattCruikshank wrote: post #386
carodgers wrote: post #17

This April 2026 paper is a fun and related read. https://arxiv.org/html/2509.24239v4 Researchers asked frontier models to play chess. Have a look at the MAR rates in Table 3. When not explicitly told which moves were legal, no model identified legal moves at a rate better than 80%. Many asked for more illegal moves than legal moves. And even when explicitly told which moves were legal, the models continued to ask for…

What happens when you ask those same frontier models to write a chess-playing program? I feel like, this is a huge stumbling block that many people have. They'll give a model their data, and ask it questions. I vastly prefer letting the model understand the schema, and then writing functions or programs to answer those questions. I feel like I get way, way better answers. I can have it write unit tests for those func…

It's a great test of cognitive abilities. There are many claims that current LLMs surpass humans in cognitive abilities so it's noteworthy that they underperform on that test.

Letting the model execute a chess program (that it wrote) would make sense if you're measuring its economic potential, but for cognition that would be cheating just like if you let a human run a chess program. The fact that the human would have a much harder time writing a useful program is irrelevant.

Re: Why I'm still bearish on LLMs after Navier-Stokes

#418

Post by zeroonetwothree »

nutrientharvest wrote: post #295
joefourier wrote: post #31

Earlier quoted context omitted.

> current frontier models > Gemini 2.5 Pro, O3, Claude Sonnet 3.7 and ChatGPT 4.1 The gap in capabilities between those models which they tested, and actual current frontier ones is enormous. I would not trust that any conclusions they made are applicable.

"Transatlantic flight will never be commercially viable, we conclude based on careful study of several aircraft designs from the 1920s"

Technology keeps advancing in a domain until suddenly it doesn’t. Where are my flying cars?

Re: Why I'm still bearish on LLMs after Navier-Stokes

#419

Post by krapp »

zeroonetwothree wrote: post #418
nutrientharvest wrote: post #295

Earlier quoted context omitted.

"Transatlantic flight will never be commercially viable, we conclude based on careful study of several aircraft designs from the 1920s"

Technology keeps advancing in a domain until suddenly it doesn’t. Where are my flying cars?

They're called helicopters.

Re: Why I'm still bearish on LLMs after Navier-Stokes

#420

Post by MattCruikshank »

cbolton wrote: post #417
MattCruikshank wrote: post #386

Earlier quoted context omitted.

What happens when you ask those same frontier models to write a chess-playing program? I feel like, this is a huge stumbling block that many people have. They'll give a model their data, and ask it questions. I vastly prefer letting the model understand the schema, and then writing functions or programs to answer those questions. I feel like I get way, way better answers. I can have it write unit tests for those func…

It's a great test of cognitive abilities. There are many claims that current LLMs surpass humans in cognitive abilities so it's noteworthy that they underperform on that test. Letting the model execute a chess program (that it wrote) would make sense if you're measuring its economic potential, but for cognition that would be cheating just like if you let a human run a chess program. The fact that the human would have…

> The fact that the human would have a much harder time writing a useful program is irrelevant.

Why?

There's a box.

You give it a problem, and it comes up with a solution.

Why does it matter to you if the box is strictly a LLM, or if the LLM can write code that it executes?

Even neater if the box is self-contained with a local model. You provide electricity, and it comes up with solutions. Why does it matter if it can do chess "in its head", or if it has to use scratch paper?

Post reply on HN