News for hackers. Stuff that matters.
Re:Sharing AI progress in mathematics by WheelsAtLarge on Tuesday October 06, 2026 @11:18PM (#49987681)
I have very little understanding of higher math, so I ask you: Was the proof due to a type of brute-force solution that could be solved had you gained enough information from reading others' work, or was it more like a proof that was sparked by an insight that came once a clue on how to solve it was put forward? I guess my question is: Was the problem proven by using a collection of everyone's work, or was it due to a brand-new insight?
Re:Sharing AI progress in mathematics by jboggan on Tuesday October 06, 2026 @11:51PM (#49987920)
I'm still digesting the proof and translating a bit from the dual case back to the primal in which I most commonly thought about it. I don't think it was a brute force proof in the sense that it combined every possible paper and commentary. It's rather odd because I feel like most of the work on the conjecture was focused on an induction proof based around graph reductions, and this proof avoided those issues entirely by offering a concrete constructive proof of finding a Hamiltonian cycle. Rather, it explicitly selected the edges not in the Hamiltonian cycle, which is in line with previous attempts via the dual.

The "aha" insight for this is actually f**ing wild, it involves a complex valued exponential sum on the edges. I've seen a lot of clever counting arguments before in graph theory but this is the first time I've seen complex roots and annihilating terms like this, the symbolic manipulation tricks in this look like things out of quantum physics. I don't understand where this trick originated, I need to really digest this.

Re:Sharing AI progress in mathematics by anilgulecha on Tuesday October 06, 2026 @11:58PM (#49987973)
BTW, reading your last paragraph reminds me of how Lee Sedol felt after move 37.
Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @12:08AM (#49988052)
Ironic, as I remember staying late at the Google office to watch that match live. I didn't really understand anything going on but I knew enough to be excited. What a decade.
Re:Sharing AI progress in mathematics by girvo on Wednesday October 07, 2026 @03:41AM (#49989563)
And we’re only a bit more than halfway through this current one. Exciting/terrifying.
Re:Sharing AI progress in mathematics by gerwitz on Wednesday October 07, 2026 @06:52AM (#49990950)
I just revisited this to make that exact comment.

I'm sympathetic to the mathematicians who are worried about the future of their field, but as an outsider I wonder if they couldn't learn from the go community's "recovery" after the introduction of an alien intelligence.

Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @12:07PM (#49994752)
Look, I quit Google a decade ago and tried to make a ChatGPT-lite LLM in my living room (turns out 2017 and GTX1080ti era was a shade too early). I knew that this technology was eventually going to revolutionize programming and mathematics and everything else. I am still flummoxed on a daily basis watching it transpire.

But also I am excited to be living through this new era of programming and new era of mathematics. I'm still saddened that I couldn't be the one to solve this old problem, but now I realize that my personal approaches were really solving a level of this problem even stronger than the original conjecture, and I'm energized to tackle those (in my free time between being a solo founder and father of 3, etc.).

Re:Sharing AI progress in mathematics by gerwitz on Thursday October 08, 2026 @02:02AM (#50002259)
Thank you for sharing your reactions to this shift; I hope it inspires many others.
Re:Sharing AI progress in mathematics by bigwheels on Wednesday October 07, 2026 @07:16PM (#50000061)
Re:Sharing AI progress in mathematics by groceryheist on Wednesday October 07, 2026 @12:07AM (#49988040)
WOW
Re:Sharing AI progress in mathematics by bamboozled on Wednesday October 07, 2026 @01:27AM (#49988614)
Why would the trick have any "origins", isn't this model creating new techniques never before seen or imagined?
Re:Sharing AI progress in mathematics by hasley on Wednesday October 07, 2026 @01:53AM (#49988816)
There is a chance that someone from a completely different field came up with a solution for a tiny part of your problem.

If you can remember the content of any scientific publication and any book in the world, you are able to make use of this knowledge in every step of you proof.

However, this does now answer how the model came up with the specific route it has taken for the proof.

Re:Sharing AI progress in mathematics by WarmWash on Wednesday October 07, 2026 @02:59AM (#49989299)
LLMs don't have super memory like that. I mean I don't know what this internal OAI model is, but at least for other LLMs, they aren't databases of training data with a smart search on top.
Re:Sharing AI progress in mathematics by hasley on Wednesday October 07, 2026 @03:14AM (#49989392)
I did not mean to say that an LLM knows literally all the publications. But the abstract knowledge is probably encoded in the weights.
Re:Sharing AI progress in mathematics by matusp on Wednesday October 07, 2026 @05:03AM (#49990116)
The agents here very likely used search. On top of that, they have boundless patience and can quickly process top K hits to find what they need. This is exactly the skill that is super useful for finding various niche sub-proofs that can help you build the final proof. A human mathematician is not going to digest 1000 papers from a different sub-field to find the needle they want, not knowing if it is actually there. AI can do it in few hours.
Re:Sharing AI progress in mathematics by AIblemblio on Wednesday October 07, 2026 @05:09AM (#49990162)
No but they have training data which teaches them certain amount of complex understandings and just not math but also physics. So this is one huge advantage.

And then they are for sure able to fill their context based on 'smart search on top' to actually progress further.

Re:Sharing AI progress in mathematics by _zoltan_ on Wednesday October 07, 2026 @05:58AM (#49990521)
As Terry Tao said, LLMs are not outsmarting us, they are out remembering us.

I'm fairly sure your understanding is not fully accurate.

Re:Sharing AI progress in mathematics by shinywth23 on Wednesday October 07, 2026 @07:01AM (#49991021)
I'm not convinced anyone really understands the difference.
Re:Sharing AI progress in mathematics by komali2 on Wednesday October 07, 2026 @01:41AM (#49988722)
As I understand it it's undetermined yet whether LLMs can actually come up with anything novel or are instead pulling from their incredibly deep corpus of knowledge to present solutions that were there but we didn't realize it because our brains aren't libraries of almost all human writing.
Re:Sharing AI progress in mathematics by TeMPOraL on Wednesday October 07, 2026 @02:46AM (#49989202)
Only in the same sense it's not yet determined about humans, either.
Re:Sharing AI progress in mathematics by RandomLensman on Wednesday October 07, 2026 @03:55AM (#49989653)
Not so sure. Was everything already "there" before humans existed?
Re:Sharing AI progress in mathematics by TeMPOraL on Wednesday October 07, 2026 @05:32AM (#49990331)
In some form and shape, yes. Humanity's creativity is a lot of marginal copying and remixing.

But obviously, it adds up to something greater than went in; in aggregate, our contributions are something to awe.

But my point is, if you zoom in at the marginal, incremental contributions of any individual human in this process, it's really hard for me to say LLMs are not at the same level already.

On this topic, people like to compare LLMs to Einstein, but as far as I know, Einstein did not zero-shot special relativity in an afternoon. He built it up incrementally over time, it took him three times longer than the time between first ChatGPT release and today, and it depended on centuries of prior art, culminating in the right observation and right notation being available to him in his moment of greatness.

Re:Sharing AI progress in mathematics by RandomLensman on Wednesday October 07, 2026 @06:10AM (#49990631)
Unless everything was there before humans existed humans created some ideas etc from scratch and not just remixed and copied.

At what level LLMs are is then an entirely separate discussion, I think.

Re:Sharing AI progress in mathematics by TeMPOraL on Wednesday October 07, 2026 @06:46AM (#49990918)
> humans created some ideas etc from scratch and not just remixed and copied.

Name three.

Re:Sharing AI progress in mathematics by RandomLensman on Wednesday October 07, 2026 @06:58AM (#49990999)
What would you accept as evidence there? Are, for example, the first names/words for colors from scratch?

So your view is that everything was there at the creation of the universe (it's a possible view, of course)? Or are there any "things" that can create ideas from scratch?

Re:Sharing AI progress in mathematics by pka on Wednesday October 07, 2026 @07:59AM (#49991525)
Recently I watched a documentary on the tanzanian Hadza tribe, one of the last hunter gatherer tribes on Earth. Their language is a distinct click and pop language and they regularly imitate animal calls (monkeys, baboons, birds) when they hunt but also when they communicate with each other, tell stories etc.

I think it's not impossible that words evolved as adaptations of the environmental sounds with which our ancestors lived. The human creativity producing DNA is also a remix of preexisting molecules formed under evolutionary pressure, so the view that it's turtles all the way down, unintuitive as it is, may not be so indefensible after all.

Re:Sharing AI progress in mathematics by komali2 on Wednesday October 07, 2026 @09:30AM (#49992596)
But LLMs aren't turtles all the way down, they stop at vector embedded tokenized words.
Re:Sharing AI progress in mathematics by TeMPOraL on Wednesday October 07, 2026 @01:55PM (#49996357)
I mean what is your criterion on invention here? On the one hand, each specific word could be seen as a new invention. On the other hand, all languages basically correlate strongly with the environment of their users - it's why LLMs turn out to be universal translators - and pattern-matching is hardly an invention, isn't it?
Re:Sharing AI progress in mathematics by pka on Thursday October 08, 2026 @05:40AM (#50003748)
I don't know, but my current intuition is that invention = more or less preexisting ingredients + some tiny spark of creativity ("connecting the dots") rather than this grand myth of "creating something from nothing". And then the tiny spark could itself be a remix of previous tiny sparks (connecting the dots between dot connections, a bit like maps between n, n+1, n+2, ...-dimensional manifolds), and so on.

> and pattern-matching is hardly an invention

Exactly. That's what I meant with turtles all the way down, until the tiny spark disappears into a dot and it turns out that creativity is just an illusion - akin to how nobody would call the gene (the spark, the pattern matching) formed by evolution (the brain connectome) "creative".

Re:Sharing AI progress in mathematics by TeMPOraL on Wednesday October 07, 2026 @01:57PM (#49996388)
My view is that LLMs meet the standards by which we judge human creativity/inventiveness, and thus that one cannot claim LLMs "just repeat, never invent" without the same being true about humans.
Re:Sharing AI progress in mathematics by komali2 on Wednesday October 07, 2026 @09:29AM (#49992569)
Pornography, "I Want it That Way" by the Backstreet Boys, torque wrench.
Re:Sharing AI progress in mathematics by throwawayk7h on Wednesday October 07, 2026 @02:04AM (#49988888)
Synthetic data allows them to train well past the limits of human writing.
Re:Sharing AI progress in mathematics by komali2 on Wednesday October 07, 2026 @09:31AM (#49992600)
What's an example of synthetic data?
Re:Sharing AI progress in mathematics by AIblemblio on Wednesday October 07, 2026 @05:11AM (#49990181)
No this is not an issue. As long as their is a way of verifying things, they do the same thing with creating novel things as humans: Searching through an infinite space of possibilities opitmized by knowledge.

They combine things, verify it and if it works and progresses the problem, they created something new.

Re:Sharing AI progress in mathematics by kelseyfrog on Wednesday October 07, 2026 @01:39AM (#49988717)
Let me introduce you to 'obscure Russian mathematicians'.
Re:Sharing AI progress in mathematics by thomasahle on Wednesday October 07, 2026 @04:10AM (#49989754)
You should try asking an LLM to look for previous papers using similar ideas. The current/frontier generation of math AI is unfortunately very bad at citing the relevant literature for techniques its using.

I asked GPT here: https://chatgpt.com/share/6ac5fd7d-0390-83ed-a02a-6d80fc64f6... and it says:

> the exact Barnette argument appears quite novel, but nearly every ingredient in its cancellation trick has a recognizable ancestor.

> The closest precedent is much closer than I expected: in fully packed O(n) loop models, people have been assigning complex phases to the two orientations of a loop and making them cancel for decades. At n=0, the phases are literally +I and -I. And the n->0 limit has specifically been used to extract Hamiltonian cycles/walks.

You can judge better than me. But it's definitely worth it having a research assistant AI with you when reading these papers.

Re:Sharing AI progress in mathematics by pbronez on Wednesday October 07, 2026 @11:36AM (#49994339)
So much about LLMs can be framed as Information Retrieval, Compression, and Search. Computers have always been good at ruthlessly hammering through a huge but finite set of possibilities. The wild thing now is that you can define that set of possibilities as "all the ideas ever published in mathematics journals."

It makes solving advanced math problems feel like cracking a hash. If it's possible, it's just a matter of compute time.

Re:Sharing AI progress in mathematics by nihalani on Thursday October 08, 2026 @01:35AM (#50002093)
Plus one to this comment so much a scaling laws, test time compute and LLM architecture decisions can be traced back/understood if you view LLMs as efficient compression databases that search based on an a retrieval query. Just turns out that they are finding super novel compression techniques that solve unknown answer
Re:Sharing AI progress in mathematics by intalentive on Wednesday October 07, 2026 @03:09PM (#49997414)
Complex roots and annihilating terms -- is it something like the derivation of Fourier / Laplace transform?
Re:Sharing AI progress in mathematics by WheelsAtLarge on Wednesday October 07, 2026 @09:52PM (#50001095)
Amazing, my first thought was these proofs came to be because there was enough information to collect, and it was just a matter of putting the data together and working the proof. It seems more than that. Your description brings to mind that, in the past, we've had discoveries that came to be because two distinct fields came together and created something new. The MRI, medicine and physics, comes to mind. Given AI's ability to juggle the world's knowledge, I get the feeling that we are bound to see many more WOW moments in the future. Thank you for your answer.
Re:Sharing AI progress in mathematics by derangedHorse on Wednesday October 07, 2026 @12:31AM (#49988200)
> Was the problem proven by using a collection of everyone's work, or was it due to a brand-new insight?

Loaded question. A "brand-new insight" is still built off the work of others. A possibly better way to frame it would be in how many subjectively unintuitive logical leaps have been made from prior work.

Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @01:32AM (#49988654)
From my current understanding (and a lot of theoretical physics I'm having to Google because the sentences I'm reading from Fable's analysis are so bizarre I think they are hallucinations) there are possibly 3 neat symbolic tricks borrowed from theoretical physics that make the heart of this proof. Forgive me for posting LLM output but I find this darkly hilarious:

"it's a matrix-tree cancellation wearing Kasteleyn's planar signs, run as a Witten index over Penrose-lineage states, evaluated as a fugacity-zero loop gas in an infinitesimal magnetic field — and the reason it reads like physics is that every one of those tools was built for partition functions"

I thought this was pure slop when I read it but there are some clear analogues in these other areas of physics, really neat computational tricks, and a very interesting paper by Penrose calculating Tait colorings I never knew about previously (extremely relevant, actually related to a separate approach I had once taken on this problem). The problem is that the paper isn't saying "aha, we were inspired by the related problems of pairing excited states and creating spanning trees out of cancelled coefficients" it just defines the function apropos of nothing. Which is kind of like the Jacobian counterexample in that it works but doesn't really explain how exactly it got there.

I really think the load-bearing concept here is "prior work". If prior work is considered papers on this problem or graph theory, yes this has one huge subjectively unintuitive logical leap. If "prior work" is the entire corpus of neat computational tricks that physicists derived to make their equations spit out something other than zero or infinity, maybe it's not so crazy?

Re:Sharing AI progress in mathematics by NitpickLawyer on Wednesday October 07, 2026 @04:04AM (#49989721)
I don't have much to add to the math parts, but I've read all your answers in this thread and wanted to thank you for taking the time to offer a detailed perspective from a subject matter expert. Thank you!
Re:Sharing AI progress in mathematics by aswegs8 on Wednesday October 07, 2026 @03:35AM (#49989520)
Actually reminds me of patent law. Prior art ist a defined term which includes all standard literature on one topic. To evaluate, whether the new solution is really inventive and thus patentable, one consults prior art, selects the most promising starting point, and from there asks oneself if an all-knowing but uncreative specialist would come up with the solution by himself. If he wouldn't, the condition of inventiveness is satisfied.

Makes me wonder how the patent space will be disrupted when that inventiveness step becomes obsolete because of LLMs. Given your example above, it seems like a combination of different methods from many different sources. This would be regarded as inventive, clearly. If eligible patents can now be brute-forced, the bottleneck becomes only selecting the most promising ones and paying for the patent.

Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @12:10PM (#49994795)
Oh man, we should talk. I have been working on a patent with ChatGPT specifically to get around two complementary patents that are now together because of a corporate merger this year. I am not sure how much longer anything is going to be patentable with this kind of design assistance available to everyone.

Also, once upon a time I wanted to be a patent lawyer. It's incredibly hard to sit for the patent bar if you have a pure math degree and don't have an engineering degree. Thankfully New Hampshire lets anyone sit for the FE exam.

Re:Sharing AI progress in mathematics by LarsDu88 on Wednesday October 07, 2026 @02:12AM (#49988950)
Did anyone else wince at seeing the phrase "load-bearing"?
Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @02:27AM (#49989075)
I did as I wrote it. I actually used that phrase often before it became an LLM-ism, just like how I rather enjoyed peppering my writing with em-dashes. Oh well.
Re:Sharing AI progress in mathematics by adamrezich on Wednesday October 07, 2026 @02:43AM (#49989180)
Language constructs becoming aggressively passé due to AI saturation is one of the craziest outcomes of all of this stuff—one which I don't think anyone saw coming.

Are there no loads left to be borne?

Re:Sharing AI progress in mathematics by RugnirViking on Wednesday October 07, 2026 @04:21AM (#49989830)
one hopes at least that the taboo on the bearing of loads is restricted to metaphorical loads only, lest lorry drivers and porters become the next victim of the algospeak spectre
Re:Sharing AI progress in mathematics by selimthegrim on Wednesday October 07, 2026 @09:56AM (#49992981)
Kasteleyn signs definitely have math counterparts (Arf invariants). They’re just not as well-known.
Re:Sharing AI progress in mathematics by jboggan on Wednesday October 07, 2026 @12:11PM (#49994807)
Right. And I'm kicking myself for not having the mathematical breadth to know about them.
Re:Sharing AI progress in mathematics by selimthegrim on Wednesday October 07, 2026 @03:18PM (#49997521)
Why? Physics people I talked to didn't know either.
Re:Sharing AI progress in mathematics by swalsh on Wednesday October 07, 2026 @09:39AM (#49992735)
It's a shame OpenAI will never publish the trace that led to the insight.