>Navier-Stokes - a so-called "Millenium Problem" - solved>literally nothing has changed>except for few people around the world losing their taxpayer-funded hobbyWhy do mathematicians and scientists lie so much about their importance?
>>17070442Have you been living under a rock? They stole the original guys's work right after it was published and even issued a public apology
>>17070442>engineered consumer products are designed, tested, certified, and shipped in a day
>>17070442The issue is how resource intensive the computation can be for some upcoming problems. We'd have to weigh the benefits of whether we want to compute it or let the mathematicians do it.Other than that, there's nothing wrong with brute forcing some problems so long as the AI has good documentation of the process to reaching that solution. Entirely new fields of mathematics have been discovered on the way to solving really hard problems, and non-spaz mathematicians are less worried about the problems being solved rather than the stuff between point A and point B possibly being interesting. For example, the results of the failed attempts to solve Fermat's Last Theorem produced far more interesting stuff than the theorem itself in my opinion.
>>17070452This really isn't as much of a problem as you may think because power is wasted all the time on extremely menial shit.
>>17070455I meant as an ideal, in case we're computing thousands of problems concurrently, the cost, power, and datacenter-sharing factors add up fast. It never hurts to optimize when we can!
>>17070458Yeah I get that, I just don't think power or cost is that much of an issue. It's been known for decades that funding research is incredibly lucrative for governments, I think I've read in a paper once of a 13x economic return on R&D investment and another finding a 17% annual economic return on publicly funded research. These LLMs will likely increase that significantly
>>17070464Yeah but I don't want an entire state's water table fucked up either.
>>17070447Adding that guy's name to the paper was a courtesy. And if you genuinely still believe in that nonsense after OAI's internal model put out results on these 700 problems you're delusional. Yeah I bet all of those were on the brink of publication as well
>>17070467>courtesytotally normal behaviour of someone not caught red handed right?
>>17070502Where are the 700+ mathematicians complaining about having their work stolen in the latest drop by OAI? Come on anon, you're smarter than this ;)
>>17070505have (you) read all 700 of them?
>>17070559Well the burden is on you. Go on and post 700 more such claims if theft, I'll wait. You can take your time, I bet they'll show up within the next month for sure
>>17070574lookThe 1 high profile problem they openai claimed to have solved was clearly stolen.Exceptions were already found in the small fry problems and yet you want to claim that the burden of proof isnt on the theives with a track record?
>>17070592The burden is on whoever claims it's stolen
>>17070595i dont claim they cant do amazing things (they certainly can) and theyll only get better in the future but in their present state. Their best(and only worthwhile) papers tend to have a track record of straight up being stolen from chats people trusted openai enough to have on them. As on right now for this specific field its clearly still stolen slop. Now show me something(other than babby tier low hanging fruit slop math) that isnt stolen.If a guy looks over to his neighbour's answers on his finals what's to say he isnt doing the same for quizzes?
>>17070447>They stole the original guys's workYou mean the work he did using ChatGPT?
>>17070442Because they solved it wrong. It was allowed by the statement of the problem, but it's not why it was interesting.
>>17070617>Now show me something(other than babby tier low hanging fruit slop math) that isnt stolen.As I said, your point of view is that all of these breakthroughs were coming anyway and thus stolen, so it's just a matter of time for someone to come out and claim theft, and the rest will turn out to contain errors.So just let it rest and wait a few weeks, /sci/ is luckily a slow board anyway. You'll look like a retard but if that's the path you want to choose, then so be it.I should also mention that there is no track record of results being "stolen," especially the NS affair can hardly be considered theft. What OAI did is add the guy's name to the paper as a courtesy.The proofs address different results and use significantly different constructions, if there was "theft," it is at best to show the model what approach doesn't work.At any rate, the proofs in the NS paper are all original and not low-hanging fruits either.Maybe if you'd read less Xitter you'd know the difference between Euler and NS, or know that viscosity wasn't even addressed in Buckmeister-Alpöge. Believe it or not, hypodissipative NS has no relation to the Millennium problem... might as well claim they stole from Cordoba and Martinez-Zoroa too lmfao any kind of contribution in the area just entitles you to the Millennium prize now I guess. Oh and let me also mention that Buckmeister NEVER accused OpenAI of theft, plagiarism, or anything really. This is all just your low-iq reading comprehension at work
>>17070645Please elaborate on what exactly about the NS problem statement was interesting and how OpenAI's proof does not add to that
>>17070442twistin and twirlin all over the nsp
>>17070447he still hasnt published his work but if it really is based off of what he did he should get partial credit
>>17070447>They stole the original guys's workStill literally no evidence btw.
>>17070844I've always pitied people who can't stand up to authority, even on an anonymous forum or in their own imagination.Despite the fact that governments and large corporations constantly lie, cheat, and steal, you can't even imagine simply acknowledging that fact instead of automatically defending them by any means necessary.
>>17070697It's the general principle, there's physical problem and then there's mathematical abstraction. The Navier Stokes paper appears to have tackled the abstraction in a way that's not relevant from the physical perspective.
I'm not anti-AI but has it been verified?
>>17070928no
>>17070928It will take a couple years to verify, and in that time clankers will have already released 6 gorillion new proofs. There are literally not enough mathematicians in the world to sift through all this data. In a couple of years, by the time they finish analyzing the 2026 Navier-Stokes solution, a single kid in his basement will be able to churn out ~8 similar papers per day. It has never been more over. Mathematicians are shitting and pissing themselves, not because they’ve been made obsolete, but because the workload and seminars required to digests these papers is already impossibly behind.
>>17070928The current view is that there are two things, the proof written in words and the mechanical proof. They're not the same, clanker fucked up.
>>17070944Why a couple of years, hasnt the blowup solution been done for the simplified version, is it that much of a leap?
>>17070502>I used AI to get my results>Someone else used AI to get better results >REEEE ITS MINE>ALL.MINE>wdym you'll give me the million dollar prize and take the credit>Fuck you! It was me, all me!I'm convinced mathematicians have unfixable irreparably large egos and would benefit from psychiatric institutionalization
>>17070592How the fuck can it be stolen? These mathematicians uploaded their work into AI for help. When AI does the helping labor, it's no longer solely the credit of the human. How about you tell me why these fags were uploading their critically important work into chatGPT.
>>17070442It's like frogger. You need to successfully move to the next lilypad to advance scienceSometimes that hop doesn't matter that much until you hop once more, but you can't make that second hop without the first hop
>>17070894That's not evidence.
Because it's a new church and they are the new clergy and monks
>I hate liars>which is why I spend weeks upon weeks on 4chan shilling a product that is created by pathological liars (californian technophiles)>who publish hundreds of "papers" without peer review>who have been caught lying and stealing>but everyone who is not part of their cabal are the REAL liars>oh also they didn't actually solve NS but that doesn't countThe mental gymnastics get more and more dire. "But slop can do X, and sock puppet bot account on reddit said it would never! Therefor I win" is the only answer this post will get
>>17070951Somebody answer this please
Somebody browses /sci/ and got triggered...
>>17070442Everything that AI is solving currently will be used by the next generation of AI to create things that seem to be magic.
