Live data from Hacker News

Erdos 281 solved with ChatGPT 5.2 Pro

twitter.com

101–110 of 310 posts

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#101
post #62
post #24

Earlier quoted context omitted.

I call it matching. Pattern matching had a different meaning.

what are you referring to? LLMs are neural networks at their core and the most simple versions of neural networks are all about reproducing patterns observed during training

You need to understand the difference between general matching and pattern matching. Maybe should have read more older AI books. A LLM is a general fuzzy matcher. A pattern matcher is an exact matcher using an abstract language, the "pattern". A general matcher uses a distance function instead, no pattern needed.

Ie you want to find a subimage in a big image, possibly rotated, scaled, tilted, distorted, with noise. You cannot do that with a pattern matcher, but you can do that with a matcher, such as a fuzzy matcher, a LLM.

You want to find a go position on a go board. A LLM is perfect for that, because you don't need to come up with a special language to describe go positions (older chess programs did that), you just train the model if that position is good or bad, and this can be fully automated via existing literature and later by playing against itself. You train the matcher not via patterns but a function (win or loose).

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#102
post #45

Earlier quoted context omitted.

Maybe it was in the training set.

I think that was Tao's point, that the new proof was not just read out of the training set.

The model has multiple layers of mechanisms to prevent carbon copy output of the training data.

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#103
post #97

Earlier quoted context omitted.

It's not. Evidence shows otherwise: Despite the "20x" length, many people actually missed the point.

I definitely missed the point because of the length, and only realized after I read replies to your comment.

Next time I'll write something shorter, or if you don't believe I wrote it... then I'll tell the AI to write something shorter.

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#104
post #47

Earlier quoted context omitted.

I suspect this is AI generated, but it’s quite high quality, and doesn’t have any of the telltale signs that most AI generated content does. How did you generate this? It’s great.

Your intuition on AI is out of date by about 6 months. Those telltale signs no longer exist. It wasn't AI generated. But if it was, there is currently no way for anyone to tell the difference.

I’m confused by this. I still see this kind of phrasing in LLM generated content, even as recent as last week (using Gemini, if that matters). Are you saying that LLMs do not generate text like this, or that it’s now possible to get text that doesn’t contain the telltale “its not X, it’s Y”?

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#105
post #50

Earlier quoted context omitted.

Pity that HN's ability to detect sarcasm is as robust as that of a sentiment analysis model using keyword-matching.

The problem is more that it's an LLM-generated comment that's about 20x as long as it needed to be to get the point across.

It's obviously not LLM-generated.

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#106

how did they do it? Was a human using the chat interface? Did they just type out the problem and immediately on the first reply received a complete solution (one-shot) or what was the human's role? What was ChatGPT's thinking time?

Heres the chat https://chatgpt.com/share/696ac45b-70d8-8003-9ca4-320151e081...

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#107

Earlier quoted context omitted.

The problem is more that it's an LLM-generated comment that's about 20x as long as it needed to be to get the point across.

It's obviously not LLM-generated.

Phew. This is a relief, honestly!

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#108
post #97

Earlier quoted context omitted.

It's not. Evidence shows otherwise: Despite the "20x" length, many people actually missed the point.

I definitely missed the point because of the length, and only realized after I read replies to your comment.

Its not just verbose—it's almost a novel. Parent either cooked and capped, or has managed to perfectly emulate the patterns this parrot is stochastically known best for. I liked the pro human vibe if anything.

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#109
Is there explainability research for this type of model application? E.g. a sparse auto encoder or something similar but more modern.

I would love to know which concepts are active in the deeper layers of the model while generating the solution.

Is there a concept of “epsilon” or “delta”?

What are their projections on each other?

Re: Erdos 281 solved with ChatGPT 5.2 Pro

#110
post #106

how did they do it? Was a human using the chat interface? Did they just type out the problem and immediately on the first reply received a complete solution (one-shot) or what was the human's role? What was ChatGPT's thinking time?

Heres the chat https://chatgpt.com/share/696ac45b-70d8-8003-9ca4-320151e081...

very interesting. ChatGPT reasoned for 41 minutes about it! Also, this was one-shot - i.e. ChatGPT produced its complete proof with a single prompt and no more replies by the human, (rather than a chat where the human further guided it.)
Post reply on HN