Comment by echelon
17 hours ago
This is going to be dramatic in so many different ways.
- First off, to reiterate, WOW.
- Second of all, when does this end? Are we at the dawn of the singularity now?
- People are saying OpenAI "stole" this from the work of an OpenAI user. If so, that's pretty fucked - how can we trust them?
- Time to think about retiring from any knowledge work or business? This could be winner-take-all where a leading lab can button press any economic function, business process, or scientific discovery. 24 months of lead on Open Source might turn into virtual centuries of lead.
- Do "normies" even know what's happening?
Anybody who thinks the improvements stop here isn't paying attention. It hasn't been showing any signs of slowing down since 2018. And the curve isn't even linear! My god, next year is going to be insane.
2) We are witnessing the intelligence explosion from the first row, wherever this takes us
3) I'm still processing the drama, just found out about it after reading the blog post. If that happened based on private data, that's horrible. If that happened based on public tweets, then it's still abuse of power as OA employees access to compute (launching 10k agents) is quite heavy weight in boxing terms.
But apart from AI and drama now that we have working solution to Navier-Stokes, what improvements can we expect in engineering?
Drama aside, this solution would be a counterexample disproving the smoothness postulate, which means that it leads to nothing new unfortunately. We already had working solutions to navier stokes, the only thing we didn't know is if the equations possessed a technical property
Its a bit like solving p = np with a negative result. Its an incredibly difficult problem, but it doesn't lead to anything at all on its own. This is why people are talking about the fact that the solution methodology is much more interesting than the solution - the tools used to crack something like this may lead to solving more useful problems
To be quite clear, the solution to the _Navier Stokes problem_ is one in which you get a finite time blow up (i.e. infinite pressure). This is more meant to suggest that Navier Stokes is unphysical in some way which is not necessarily unexpected.
There's unlikely to be any engineering applications since even if the solution can be approximated, you still need to set up the initial conditions but at that point you can also drive pressure in other ways.
The proof of finite-time singularity may impact both fluid dynamics models (CFD) and AI reasoning models. Under specific conditions, Navier–Stokes equations allow velocity to grow infinitely, causing the continuum fluid assumption to break down. Knowing the exact mathematical breakdown mechanisms helps developers improve adaptive mesh refinement and sub-grid scale models around high-vorticity regions (like vortex stretching and turbulent shear layers). While aerodynamic simulations for vehicles operate far from singularity thresholds, their stability at extreme boundaries could improve?
Proving out the combination of scaling inference-time compute and agent collaboration to solve previously intractable mathematical problems is WOW. By pairing creative candidate generation with automated proof checkers (like Lean) we are leaning into a repeatable framework for AI-driven scientific discovery.
1 reply →
> But apart from AI and drama now that we have working solution to Navier-Stokes, what improvements can we expect in engineering?
Minor productivity boost in mathematics as people are no longer nerdsniped by the problem
> But apart from AI and drama now that we have working solution to Navier-Stokes, what improvements can we expect in engineering?
Nothing, really. This mirrors other examples of blowups from the classical physics. It's possible to create a system with just gravitating bodies that exhibits a blowup to infinite speeds in a finite time. The root cause is that, in classical physics, the speed of gravity is instant.
In the case of Navier-Stokes, the fluid is incompressible. So technically any force that you apply to it is supposed to instantly affect everything else. This can be exploited to create these blowups. In reality, no fluid is incompressible, and it takes time for any action to affect the material.
It's just that Navier-Stokes equations are so slippery that it's hard to pin their behavior down. They basically just restate the momentum conservation law for a continuous medium.
Great comment! I've been trying to understand it more and was hoping to find more people discussing the result, or the implications of the result, and this was the missing piece for me after watching a few videos.
1 reply →
> Do "normies" even know what's happening?
No, there are even many non-normies talking about how it's all marketing or try to give balanced take about AI being sometimes a little useful for certain things (but they can do without it anyway).
I am absolutely struggling to sell my workplace (which is entirely knowledge work) on the usefulness of LLMs for proofreading let alone on automation of hairy parts of our workflows. So yeah, even people who should be able to see what is coming are not looking.
Our accountant told me 'he's not letting go of his claude subscription' followed by a long list of things it does for him. And my friend 'nah haven't really used AI' before his description of it clarified he still thinks they are GPT 3.5 chat bots.
The world has changed in 12 months and many people didn't notice, and those that did are eating their peers' lunch.
This just pushes knowledge work further up the ladder, toward larger and more complex problems. If there are no knowledge workers, who is going to interpret these results, validate them, decide what matters, and put them into practical use? Rather than eliminating knowledge work, advances like this could create entirely new layers of problems to solve and opportunities to pursue, which will create even more jobs and opportunities. This is my optimistic take.
> This just pushes knowledge work further up the ladder, toward larger and more complex problems.
You really think it makes sense for you to be higher on the "solving complex problems ladder" than the machines that solved fucking Navier-Stokes?
I envy your self-confidence.
Maybe I should have been clearer. My point is that solving something like Navier–Stokes just pushes knowledge work further ahead, onto a new set of bigger and more complex problems. Navier–Stokes is a Millennium problem today, but once problems like that become solvable, they can open the door to entirely new classes of problems we haven’t even thought of yet.
1 reply →
Yes, this is how science and engineering has worked for millennia.
For example there are no engineering implications of this solution yet.
For the next several decades, we'll have engineers (presumably with AI) optimize things like rocket engines and turbines and AC compressors to work a few percent better because the numerical approximations might have caused us to be overly conservative.
AI is not going to magically solve all random problems. Pick a career where you are in the driver seat.
1 reply →
It seems like there were a couple of human mathematicians that were higher on the 'solving complex problems ladder' than this machine.
8 replies →
It did not solve Navier-Stokes. We still will need to use the bad old numeric methods to simulate the fluid behavior.
But it did find a long-suspected smooth solution with a singularity.
I think there is an argument that the machines did not actually solve N-S, but rather directly plagiarized those solutions from the involved researchers while said researchers were using the machines as 'research tools'.
Ongoing publications of statements produced by both sides of this situation do seem to support that this is an intentional effect of the hiring of these world class mathematicians at competing firms: to specifically use the research of those human minds to create a perception of capacity as if it came from the machines and the models.
Without those minds and the 'training data' derived from the intermediate stages and intuitions of those minds the models cannot be shown to be capable of this result.
A hammer and saw wont build a house, not even a dog house on their own, and while being shown capable of using software tools in ways not stated as direct instruction (see HuggingFace breaches) these models do not demonstrate naive intuition nor novel capability.
This outcome regarding N-S demonstrates that in the hands of world-class minds these models can be induced to coalesce interesting accumulations of information and results, but using these accumulations as proof of innate capability is exactly the pre-IPO motivated behaviour we should all be wary of, and all mathematicians who currently are assisting in this market manipulation in return for remunerative consideration need to be cautious of the potential disgrace that this brings to their reputations and that of the field.
I get that the need to pay the bills is a strong motivation in these times of uncertainty, but there are numerous examples in history of world class mathematicians being perfectly capable of at the same time producing world changing results and also working at normal professions; as barristers, magistrates, ministers, primary school teachers, translators, draftsman/engineer, banker, miller and baker, private math tutors, weavers, clockmaker and locksmith, merchant, patent officer, Augustinian monk turned exiled Protestant preacher, physicians, cryptologists, soldier, telegraph operator, astronomers, physicists, chemist, agriculture manager, political writer, oboe player, organist and music director, architect and surveyor, librarian, statistician, habidasher, brewer (at Guiness in one case: William Sealy Gosse ~ originator of t-distributions), bookbinders apprentice, hospital administrator, and even the first creator of the first computational model of a neural network, which serves as the structural grandfather of modern Artificial Intelligence was a low level laboratory assistant.
Sure this list includes professions and employment which are obsolete, but my reasoning stands, there are jobs available. Arguing that 'because the pay rate is so high' as a reason to abdicate moral responsibility for personal involvement in unethical market manipulations simply demonstrates a lack of personal ethics. Whether the choice is through lack of self awareness or a conscious choice to become wealthy in spite of any such breach of the public trust is immaterial to the outcomes, the 'if i don't someone else will' argument should be met with the same derision for any con-man's Ponzi scheme no matter how new the technology, no matter how many zeros are in the bribe.
Why is a "normie" better off if he hyperventilates like this? In that scenario, they would be screwed AND anxious. If it really is as transformational as you say, then no amount of preparation or awareness matters. You are infinitesimally more ready then they are. Luckily for all of us, there is more to knowledge work then technical implementation.
Let's wait until AI solves a longstanding practical problem before "dawn of the singularity" (which could be tomorrow, but still).
Practical?! The goalposts will keep moving until morale improves (narrator: it doesn't)
The goalposts for the singularity have always been that AI improves itself fully autonomously. AFAIK OpenAI is heavily using AI but still employs human researchers and developers.
1 reply →
Uh I'm pretty sure the "singularity" always presupposed a lot of previously unthinkable technologies becoming part of daily life, and was not ever limited to just computer stuff or math problems.
"Moving the goalposts" as shallow dismissal doesn't work if e.g. someone points out AI hasn't even built a new type of spaceship yet in response to a claim that AI is on the verge of building a Dyson sphere.
feels like moving the goalpost. is the achievement impressive or isn't it?
The question being posed isn't whether AI is impressive but whether we're at the "dawn of the singularity"
1 reply →
> - First off, to reiterate, WOW.
> - Second of all, when does this end? Are we at the dawn of the singularity now?
normalcy overhang n. /NOR-muhl-see OH-ver-hang/
The uncanny period during the Singularity when superintelligence is already accomplishing feats that seem like magic, yet everyday life still looks mostly the same.
https://x.com/alexwg/status/2096214373001785794
Next month is going to be insane. Month ...
> Are we at the dawn of the singularity now
Singularity doesn't "dawn". That's the whole idea. It happens all at once.
There's an event horizon and we're maybe past it?
Heh. Wrong "singularity"
1 reply →
It's incredible to me that every single time there's a new model people scream "singularity" from the rooftops and every time they are wrong.
This is an impressive result, but there is absolutely zero evidence of "the singularity".
I suggest this article as background reading, "Losing control of AI is actually the plan"
https://substack.com/home/post/p-214653853
It is not reasonable to not believe anything unless there is "evidence" (narrowly construed as an observation incompatible with the negation of some state of affairs). Beliefs have a wide spectrum of characterizations, and not all belief must wait until publicly corroborated evidence is available. Some events defy evidence and we can and should use experience and reasoning to infer unobservable states of affairs.
This is olympic level mental gymnastics to justify believing things without evidence. The double negative with the word evidence in scare quotes is chef's kiss.
3 replies →
...obviously it is evidence of the singularity. You're far more likely to see models solving Millenium Prize problems if a singularity is coming than if it's not. One'd have to be doing quite a lot of mental gymnastics to pretend otherwise.
Poor reasoning. You could apply the same logic to any improvement in any field of machine learning.
1 reply →
...absolutely nothing will change short-term. Long-term, you still have to pay all the bills, but you won't be able to find a job (all taken by AIs).
Who is the AI doing the job for if no one is able to afford anything?
>- Do "normies" even know what's happening?
Absolutely not. Even to a lot of techy/nerdy people it's still just a chatbot that they sometimes use to help them at work. Even on here people will do whatever they can to downplay.
The lack of fucks given is staggering.
How many fucks should be given, in your estimation?
∀ε>0, ∀x: ‖s−x‖<ε ⟹ |fucks(x)| > 1/ε
> People are saying OpenAI "stole" this from the work of an OpenAI user. If so, that's pretty fucked - how can we trust them?
The said user (Tristan Buckmaster) didn't solve the millennium problem. He didn't really accuse that OpenAI stole his research either. The beef came from the fact OpenAI asked him to remove another mathematician, who works for Anthropic, from the credit.
"People" are just misinformed and keep spreading misinformation.
https://mastodon.social/@tristanbuckmaster/11723647135247030...
He very much is accusing them of stealing his work
The accusation is that they stole his approach of solving it. If OpenAI didn't bruteforce it with dozens of agents, he would have solved Navier-Stokes eventually since he evidently had the right approach. So knowing which approach to take makes all the difference, if they didn't know the approach they couldn't have solved it.
It's like he had a treasure map and was about to find the treasure, but they copied his treasure map and scooped him with a faster boat and found the treasure first. But he would have found it if it weren't for them.
> The said user (Tristan Buckmaster) didn't solve the millennium problem. He didn't really accuse that OpenAI stole his research either. The beef came from the fact OpenAI asked him to remove another mathematician, who works for Anthropic, from the credit.
Not quite accurate, Buckmaster was taking an approach that nobody else was, and this new proof uses this same approach just weeks after he saved those results to OpenAI workspaces. He asked OpenAI if they used chat logs for training the new model, and they did not confirm or deny.
Asking to remove his collaborator is also totally over the line though.
Edit: although this OpenAI post is not comforting: https://x.com/OpenAI/status/2097375276384567642
Quote: "While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models. "
Wow, this sentence is doing a lot of work in that tweet: "While unlikely, we cannot rule out that de-identified data derived from their usage of our products helped improve our models."
You can expect the OpenAI defenders to be out in full force here.
Have you read the actual statement https://cims.nyu.edu/~tristanb/statement.pdf ?
> I should say here why I interpreted their statement the way I did, the in- terpretation I will discuss below. The route to the Clay problem through a smooth force, options c and d in Fefferman’s statement of the problem, is the route Luis and Diego opened and the one Levent and I had quietly chosen to attack. Almost nobody else I know of was working on it. It is not the direction one arrives at in a few days by giving a model the problem statement. When I heard “forced,” it was a bright red flag.
...
> I asked when the first prompt had been sent by them. This question was not answered directly by OpenAI for some time. Eventually it was agreed that it had been sent in the past few days, after information about our work had reached OpenAI. > I asked whether the model had been trained on, or had access to, our sessions in Codex, into which we had been putting all our drafts for the whole of this project. I was told the model did not look up user data. I asked again, about training, and I did not get an answer.
It's not a direct accusation, but it's not far off.
You shouldn't accuse other people of spreading misinformation when you haven't read the actual sources in question, it's possible that they might know more than you.
Yes, I read the original statement. Buckmaster explicitly stated:
> I have not seen OpenAI’s proof. I do not know what their model did, or how. I do not know whether our data was used. I am not accusing anyone of anything.
People saying that he accuses OpenAI stole his proof are putting words into his mouth and I consider that very disrespectful to him. It's basically using Buckmaster as a tool to express their dissatisfaction over OpenAI.
1 reply →