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

AGI is Mathematically Impossible 2: When Entropy Returns

philarchive.org

Post reply on HN
Print view View original

41–50 of 437 posts

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#41

Post by rusk »

vessenes wrote: post #32
ICBTheory wrote: post #2

This paper presents a theoretical proof that AGI systems will structurally collapse under certain semantic conditions — not due to lack of compute, but because of how entropy behaves in heavy-tailed decision spaces. The idea is called IOpenER: Information Opens, Entropy Rises. It builds on Shannon’s information theory to show that in specific problem classes (those with α ≤ 1), adding information doesn’t reduce uncer…

Thanks for this - Looking forward to reading the full paper. That said, the most obvious objection that comes to mind about the title is that … well, I feel that I’m generally intelligent, and therefore general intelligence of some sort is clearly not impossible. Can you give a short précis as to how you are distinguishing humans and the “A” in artificial?

Not the person asked, but in time honoured tradition I will venture forth that the key difference is billions of years of evolution. Innumerable blooms and culls. And a system that is vertically integrated to its core and self sustaining.

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#42

Post by Animats »

Penrose did this argument better.[1] Penrose has been making that argument for thirty years, and it played better before AI started getting good.

AI via LLMs has limitations, but they don't come from computability.

[1] https://sortingsearching.com/2021/07/18/roger-penrose-ai-ske...

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#43

Post by mort96 »

alganet wrote: post #16
ninetyninenine wrote: post #5

Without reading the paper how the heck is agi mathematically impossible if humans are possible? Unless the paper is claiming humans are mathematically impossible? I’ll read the paper but the title comes off as out of touch with reality.

What makes you think that human intelligence is based on mathematics?

I will answer under the metaphysical assumption that there is no immaterial "soul", and that the entirety of the human experience arises from material things governed by the laws of physics. If you disagree with this assumption, there is no conversation to be had.

The laws of physics can, as far as I can tell, be described using mathematics. That doesn't mean that we have a perfect mathematical model of the laws of physics yet, but I see no reason to believe that such a mathematical model shouldn't be possible. Existing models are already extremely good, and the only parts which we don't yet have essentially perfect mathematical models for yet are in areas which we don't yet have the equipment necessary to measure how the universe behaves. At no point have we encountered a sign that the universe is governed by laws which can't be expressed mathematically.

This necessarily means that everything in the universe can also be described mathematically. Since the human experience is entirely made up of material stuff governed by these mathematical laws (as per the assumption in the first paragraph), human intelligence can be described mathematically.

Now there's one possible counter to this: even if we can perfectly describe the universe using mathematics, we can't perfectly simulate those laws. Real simulations have limitations on precision, while the universe doesn't seem to. You could argue that intelligence somehow requires the universe's seemingly infinite precision, and that no finite-precision simulation could possibly give rise to intelligence. I would find that extremely weird, but I can't rule it out a priori.

I'm not a physicist, and I don't study machine intelligence, nor organic intelligence, so I may be missing something here, but this is my current view.

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#45

Post by Workaccount2 »

deadbabe wrote: post #17
like_any_other wrote: post #3

So does the human brain transcend math, or are humans not generally intelligent?

First of all, math isn’t real any more than language isn’t real. It’s an entirely human construct, so it’s possible you cannot reach AGI using mathematical means, as math might not be able to fully express it. It’s similar to how language cannot fully describe what a color is , only vague approximations and measurements. If you wanted to create the color green, you cannot do it by describing various properties, you m…

I don't think it would be unfair to accept the brain state of green as an accurate representation of green for all intents and purposes.

Similar to how "computer code" and "video game world" are the same thing. Everything in the video game world is perfectly encoded in the programming. There is nothing transcendent happening, it's two different views of the same core object.

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#47

Post by rusk »

daedrdev wrote: post #29

Clearly nature avoids this problem. So theoretically by replicating natural selection or something else in AI models, which arguably we already do, the theoretical entropy trap clearly can be avoided, we aren't even potentially decreasing entropy with AI training since doing so uses power generation which increases entropy

It can be avoided certainly, but can it be avoided with the current or near term technology about which many are saying “it’s only a matter of time”

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#48

Post by WhitneyLand »

ICBTheory wrote: post #2

This paper presents a theoretical proof that AGI systems will structurally collapse under certain semantic conditions — not due to lack of compute, but because of how entropy behaves in heavy-tailed decision spaces. The idea is called IOpenER: Information Opens, Entropy Rises. It builds on Shannon’s information theory to show that in specific problem classes (those with α ≤ 1), adding information doesn’t reduce uncer…

“This paper presents a theoretical proof that AGI systems will structurally collapse under certain semantic conditions…”

No it doesn’t.

Shannon entropy measures statistical uncertainty in data. It says nothing about whether an agent can invent new conceptual frames. Equating “frame changes” with rising entropy is a metaphor, not a theorem, so it doesn’t even make sense as a mathematical proof.

This is philosophical musing at best.

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#49

Post by pvg »

JdeBP wrote: post #36

This has a single author; is not peer-reviewed; is not published in a journal; and was self-submitted both to PhilArchive and here on Hacker News.

There's nothing wrong with any of that, for an HN submission. The paper itself could be bad but that's what the discussion thread is for - discussing the thing presented rather than its meta attributes.

Re: AGI is Mathematically Impossible 2: When Entropy Returns

#50

Post by Workaccount2 »

ImHereToVote wrote: post #13
like_any_other wrote: post #3

So does the human brain transcend math, or are humans not generally intelligent?

Humans use soul juice to connect to the understandome. Machines can't connect to the understandome because of Gödels incompleteness, they can only make relationships between tokens. Not map them to reality like we can via magic.

Stochastic parrots all the ways down

https://ai.vixra.org/pdf/2506.0065v1.pdf

Post reply on HN