Live data from Hacker News

Position: LLMs Can't Jump

openreview.net

141–150 of 233 posts

Re: Position: LLMs Can't Jump

#141

This is literally an opinion of one dude which is not backed by any kind of quantitative evidence. It's actually possible to answer this question rigorously: 1. Define a scientific result which qualifies as a "jump". They should be frequent enough that they happen every year - otherwise one might say humans can't jump either. 2. Identify all such "jumps" in articles published in 2026, and use LLM with 2025 knowledge…

You don't need any reproduction. Assuming researchers are using LLMs, you should just see the number of jumps increase as models get better.

Re: Position: LLMs Can't Jump

#142

Earlier quoted context omitted.

I agree completely - behaviorally the models have changed drastically due to RLHF, RLVR and now maybe even more so due to agentic harnesses. But the mechanism of prediction hasn’t changed, that was all I was clarifying.

What about multi-token prediction and speculative diffusion? That’s a different mechanism of prediction, even if it serves only to accelerate decoding.

As you say, that's just an efficiency play and, as I understand it, doesn't change the behavior of the models beyond perhaps a small amount of sampling noise.

Re: Position: LLMs Can't Jump

#143
Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean very different things, such as, "I love my children" or "I love apple pie". Clearly the word "love" here has a certain magnitude that is not being expressed, yet it is understood by the listener somehow.

So, I do sort of buy into this idea that Einstein was simulating the world and running experiments on those simulations in ways that were beyond what you could encode in natural language. Will AI be capable of doing this, if it is bounded by training data that is composed almost entirely on language? One might argue that if AI is training on a lossy encoding/representation of the human experience, how will it be able to simulate anything beyond that experience? Unless it does so in a way that we manage to do when we image objects beyond 3D. But now I'm just rambling.

Re: Position: LLMs Can't Jump

#144

Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean…

I think this is very critical concept and existing gap which does not makes into AI conversations. Interestingly enough a lot of Si Fi movies have captured this where the AI starts to feel and have "thoughts". Have to give kudos to the authors for being so creative and imaginative.

Re: Position: LLMs Can't Jump

#145

Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean…

Or, ask yourself the following question:

If you read every single book there is about The Grand Canyon, and watched every single video and/or documentary about The Grand Canyon, do you believe that you have fully experienced The Grand Canyon? Or do you just have to be there to fully experience it.

I dunno. Substitute in whatever you want for "The Grand Canyon". Maybe climbing Mount Everest or walking on the Moon. The point is that maybe the human experience is more vast than what is written about it.

Re: Position: LLMs Can't Jump

#146
>"While Generative AI has mastered Induction (statistical pattern matching) and is rapidly conquering Deduction (formal proof), we argue it lacks the mechanism for Abduction—the generation of novel explanatory hypotheses."

Interesting!

Induction Vs. Deduction Vs. Abduction!

(You know, if you like Logic, Philosophy, Law, or... just plain different ways to think/reason about something! :-) )

Re: Position: LLMs Can't Jump

#147

Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean…

The lossy compression of language is why we should find it unsurprising that LLMs tend to perform better at code, than at human language tasks or reasoning. While there can be subtle semantic differences in real codebases (using "null" to mean "unknown" in one context, versus "intentionally blank" in another), there is a much tighter coupling of semantics to meaning (low ambiguity) compared to "love" in English (let alone any inexpressible je ne sais quoi).

With apologies if this is common knowledge at this point, 3Blue1Brown has been doing an excellent series on compression, and its relationship to intelligence (or more controversially, that they are one and the same): https://www.youtube.com/watch?v=l6DKRf-fAAM

But that also throws in sharp relief, that there is vastly more to the human experience than intelligence alone: qualia, desire, gut instinct, intuition, emergent creativity. (Whether the "God of the Gaps" for the delta between capabilities of human vs AIs is fixed, or diminishing, or even shrinking to zero, remains an open experiment we're all living through.)

Re: Position: LLMs Can't Jump

#148

Earlier quoted context omitted.

I get really tired of Einstein being portrayed as if he’s the greatest scientist who ever lived, when he doesn’t even belong in that conversation. If Einstein never existed, the fields he was in were already moving strongly in the directions of his conclusions. If Maxwell or Newton had never existed, the world today would be quite different.

Einstein did a lot, he not only published the special and general theories of relativity, he also contributed to quantum physics and was the first to deduce the size of an atom using statistical arguments, essentially proving the existence of atoms.

Not at all.

The size of an atom was computed for the first time 40 years earlier, in 1865, by Johann Josef Loschmidt, who was an Austrian scientist.

Loschmidt had determined the value of what today is called "Avogadro's number", despite the fact that Avogadro had no idea about how one could find the value at that number. The contribution of Avogadro had been the law that ideal gases have the same number of molecules per volume when the pressure and the temperature are the same, but he did not know how that number might be determined.

Based on the results of Loschmidt, which were also used by Maxwell in his works on statistical physics, a few years later, in 1874, George Johnstone Stoney computed the value of the elementary electric charge, which he later named "electron".

Loschmidt was the one who first measured and weighed the atoms, establishing their existence beyond any reasonable doubt. Before him, many believed that the hypothesis of Dalton about the existence of atoms is just a convenient theory for explaining the laws of chemistry, but some other better explanation might be discovered in the future.

Einstein published an explanation of the Brownian motion, which verified that everything fits in the known theories and it matches the expected values. It did not contain any novel number or law, but it was a useful confirmation of the current theories.

In my opinion, the greatest achievement of Einstein has been the paper published by him in which he revealed the existence of the stimulated emission of electromagnetic waves by hot bodies, besides their well known absorption and spontaneous emission.

This paper had tremendous practical applications, leading to the discovery of masers and lasers, without which the modern technology would have never existed.

While Einstein's theory of gravitation was more intellectually challenging, its practical applications are minor, almost negligible. Moreover, even after more than a century we do not know exactly how adequate it is for describing gravitation.

What is certain is that Einstein's theory of gravitation cannot be anything else but an approximation, as it is based on averaging the distribution of matter in space, without taking into account the concentration of matter into elementary particles.

Another contribution of Einstein that is more important than his most frequently cited works is his support for Satyendra Nath Bose, which lead to the recognition of the Bose-Einstein statistics, which together with the Fermi-Dirac statistics plays an essential role in quantum physics.

Re: Position: LLMs Can't Jump

#149

Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean…

[dead]

Re: Position: LLMs Can't Jump

#150

Probably too late for this, but I have argued before that language is a fundamentally lossy encoding of the human experience. We do our best to describe what we're seeing and experiencing using language, which is fantastically expressive, but it has its limits. I think we see glimpses of this when we find ourselves saying things such as, "it's impossible to put it into words" or we overload certain words when we mean…

Or, ask yourself the following question: If you read every single book there is about The Grand Canyon, and watched every single video and/or documentary about The Grand Canyon, do you believe that you have fully experienced The Grand Canyon? Or do you just have to be there to fully experience it. I dunno. Substitute in whatever you want for "The Grand Canyon". Maybe climbing Mount Everest or walking on the Moon. The…

I think I could get very close via video. But that depends on a lot of my non-canyon experience moving around the world, and LLMs are very flawed in their ability to input video, so I think they would not get nearly as close.
Post reply on HN