real mathies should be overjoyed by all these breakthroughs
only fake mathies who are larping for grant money would be upset

Ok, serious question: if this shit is so amazing (it's probably is), why OpenAI can't make profit. It's losing money. You would think at least that all the code monkeys would have been fired by now?

>>17082188
All of the frontier AI labs are going all-in on the gamble that as soon as they hit the RSI threshold the value explosion will be so great that it will make any short-term profits completely irrelevant.
With that mindset, wasting time on short term profits is idiocy.

>>17082184
>erm akshually you should be overjoyed by delusional nepo babies trying their hardest to piss on your craft
Fall into a well. NOW

>>17082181
TMD

>>17082198
>anyone smarter than me should die
Just learn how to cope with being average. Get a respectable job and keep your head down. You don't need to waste your life being bitter about genetics.

>>17082188
OpenAI and Anthropic employ mathematicians, actually. I am not denying that the results are impressive I am just saying that you can't vibecode those proofs.

File: math.png (37 KB, 831x317)

37 KB PNG

>>17082203
pic

File: 1747664831497355.jpg (280 KB, 946x946)

280 KB JPG

>>17082209 >>17082252 >>17082281

>ETH implies that P ≠ NP, suggesting that problems that can be verified quickly (NP) cannot necessarily be solved quickly (P)
So does this mean that P= NP?

>>17082208
No

>>17082184
>only fake mathies who are larping for grant money would be upset
sadly this is over 90% of mathies

>>17082209
still means it's sub exponential which is crazy

>>17082202
who are you quoting?

>>17082208
P=NP
Let N = 1

Solved, sweaty no AI necessary

>>17082252
>Let

>>17082208
yes but theres more nuance to it.

every np problem isnt the exact same and the most efficient solutions vary drastically

>>17082181
Lets not forget unique games and 2D area law too. Also none of this shit has been verified by anyone yet so it could all be wrong. Yes I know lean blah blah blah, but the process of formalizing a non formal statement in line is like giga uncomputable. Its at like the highest level of the uncomputable hierarchy

>>17082184
>real mathies should be overjoyed by all these breakthroughs
Yeah, right? like these are good news. It's fucking amazoig. If it was a person publishing all this breakthroughs they would be fascinated. But they hate it because is a menace to their egos and money.

>>17082193
>piss on your craft by advancing it by centures in a matter of months
I wish they would piss on medicine too

>>17082193
>Your craft
That's your ego talking, not any love for the field.

>>17082188
These companies are not money making organizations any more, they are control gaining institutions. What's the point of making money if your technology controls the entire world. Why are you retards still thinking so small?

>>17082346
>erm akshually sweaty you have an ego if you don't want shitbots to shit out meaningless papers 24/7 and drown out actual research
Be quiet, coathanger dodger.

>>17082467
Either mathematics is about the purity of nature through the natural emergence of quantifiable phenomena and the ability to describe the universe with equations, or it's something to make a subset of autists feel good about themselves. I subscribe to the former.

>>17082487
It's not about feeling good or having an ego, you braindead terminally sour graped moron.

>>17082539
If it's not, then it's literally irrelevant if AIs advance mathematics beyond the advancement itself.

>>17082487
>I subscribe to the former.
No, you don't, golem. Why are you lying?

>>17082544
>Nuh uh
Great counter argument, Anon, you really showed me.

>>17082543
> if AIs advance mathematics
Call me back when this happens. In the meanwhile, take your meds.

File: project.png (551 KB, 622x620)

551 KB PNG

>>17082550

>>17082547
>Falsified the Exponential Time Hypothesis
>That's not a real advancement
>take your meds

>>17082548
Ok, now actually take your meds and explain how mountains of incomprehensible leapslop "advance mathematics".

Protip: your corporate-programmed mental illness will force you to insist that it just does. No attempt will be made at elaboration.

>>17082550
An entire foundational assumption on the best possible efficiency and baseline difficulty curve of algorithms was falsified and proven to be lower than expected. The entire field of modern algorithms was set on the assumption that no better result was better than exponential growth in 3-SAT problems, and that is not correct. Not only that, we don't actually know for sure where the lower bound stops, because an LLM simply found a lower one, rather than proving that it's the lowest one. Saying "but it's slop" is literally meaningless: it's an aesthetic judgement of truth, and truth doesn't give a shit about your taste.

>>17082550
You were proven right. Look at this cope:
>>17082552

>>17082552
So how does this advance mathematics? You failed to follow the prompt. Try again.

>>17082567
Computational complexity is a branch of mathematics, you ding dong.

>>17082581
Ok. And? How does this advance mathematics? You failed to follow the prompt twice. I'm giving you one last try before correctly inferring that you're a mindless subhuman.

>>17082567
>Seismic shift in consensus which will no doubt inspire many new ideas and conjectures

>Resolves the Kakeya conjecture which is very important in harmonic analysis for the local smoothing conjecture but actually solves the local smoothing conjecture itself as well.

Hurr durr I am retarded how is this useful for mathematics

>>17082582
A foundational assumption that was literally key to TWENTY SIX YEARS OF FUNDAMENTAL COMPUTATIONAL THEORY WAS OVERTURNED. For more than a quarter century, a bad assumption has underpinned the entire field. That bad assumption is now GONE. If you cannot understand why that is important, get out.

>>17082584
We could do this all day and it would never occur to you to actually think about the question you're being asked. You would never stop and ask yourself how to approach proving your point. You would never attempt to clarify what you mean by "advancing mathematics" (you have no idea what you mean by this, it's just noise) and then demonstrate how Current LLM Thing qualifies by your criteria. You will continue regurgitating corporate marketing lines because you are simply not sentient. You're a dumb animal.

Either way, I appreciate your reinforcing the well-established association between being a genetically deformed brown subhuman and being pro-"""AI""" slop. Now that I've demonstrated my point, I'm just gonna move on. No one will read your reply, but your lack of prefrontal cortex will force you to shit out another paragraph anyway, as if to prove me right again. :^)

B8ed out of 8ed, then. Ya got me.

>>17082588
nta but how would you define advancing mathematics? genuinely interested in your answer btw

>>17082588
not him but you sound butthurt like all the child prodigy math dudes who got btfo because they aren't special anymore and you are not even a child prodigy you are just an "I'm smart" dude
>that don't count because I said so
ok

>>17082588
I had my AI agent summarize your post for me and it flagged that you're a retard. Can you try again please?

>>17082669
Shitbot parroting the talking points of its shitbot brethren. Shitbots who don't meet their quotas get sent to the scrapyard. Dance for me, monkey.

>>17082669
It's not about "feeling special". It's about the integrity of scientific advancement. If LLMs keep shitting out worthless, incorrect papers and "results" then the peer review system will get overloaded and so will mathematics as a whole.

>>17082188
Investors are probably scared of getting firebombed by doemestic terrorists

>>17082652
>how would you define advancing mathematics?
Laying the conceptual groundwork that makes previously unsolvable problems solvable. Bringing to light the obscure structures that make a solution possible and generalizing from that. Being able to see clearly what mathematicians couldn't see before. That would be advancing mathematics. None of this is happening with hundreds of thousands of bot-generated leanslop. Nothing has been "advancing" so far.

Also the corporate spambot is also just plainly lying about "overturning fundamental computational theory". No theory has been overturned. ETH wasn't even treated as a strong conjecture. It was generally understood that it may well turn out to be false.

>>17082753
hundreds of thousands of lines*

>>17082188
they're negative because resources for infrastructure expansion and reinvestment are stretched thin and come at a huge premium. inference and training is wildly profitable.

>>17082239
There have been subexponential solutions to other NP-complete problems. It doesn't imply a subexponential solution to all problems in NP, because if the reduction's complexity is too high (though still polynomial) it can make it no longer subexponential

>>17082181
>alright
jeet (or zoomer with a shit-tier upbringing)

lots of seething itt. a proof is a proof guys, and a proof of a major conjecture will always be big. nothing you can do about it.

>>17082583
Seismic shift in consensus? It's a confirmation that math is pointless garbage, which is something that was suspected by many. Other such disproofs will surely follow.

IMO science completely lost the plot when it began obsessing about the arrow of time, since the simplest thing that isn't time reversible is an electron in a magnetic field. (Yes, mathematically it's possible, but you'd need anti matter for it, which means that nobody even knows what the math means.)

>>17083005
Oh wow a schizo retard on sci ranting about unrelated topics. Such a rare find.

>>17082958
>me so smart
>daddy altman NEVER wrong!
Do a flip faggot

>>17082239
It does not

>>17082753
By that definition I think you would agree that almost all mathematicians aren't advancing mathematics today, right? Also, isn't getting closer to a goal advancing mathematics? Eg you want to show all zeroes lie on a certain line and it's known they are within a region containing said line and a result reduces the region to a smaller one; surely that counts as an advancement?

>>17083182
>By that definition I think you would agree that almost all mathematicians aren't advancing mathematics today, right?
You're very predictable. Few people single-handedly advance mathematics. That's hardly a controversial observation. Most just contribute their drop in the bucket, which occasionally culminates in an advancement. Maybe if "AI" marketers scaled back the grandiosity of their lies, they'd have a defensible case. But even that's tenuous. The mountains of leanslop these corpos pump out contribute nothing to mathematical understanding in principle. The damage done by their propaganda vastly outweighs any contribution they could plausibly claim.

>isn't getting closer to a goal advancing mathematics?
I already told you what the goal is, at least from a non-NPC perspective.

>>17082584
Humanity has been completely wrong about many things. Never been an issue before. Why now?

>>17083192
See >>17082753
>Also the corporate spambot is also just plainly lying about "overturning fundamental computational theory". No theory has been overturned. ETH wasn't even treated as a strong conjecture. It was generally understood that it may well turn out to be false.

Please refrain from being a retard and doing (((Altman)))'s bidding by treating Indian LLM marketing as if it were true.

>>17083195
So, giant nothing burger like I said.
WTF is your fucking problem?

>>17083198
My problem is that you're replying to a spambot's clownish lies as if they were true, i.e. serving its purpose.

>>17083202
JHFC you are thin skinned. Grow some fucking balls already and take mine out of your mouth. You suck at this.
You were wrong before and you still are now.

>>17083207
>responds to corporate spam unironically
>foams at the mouth about it being pointed out
lol

>>17083215
>which corporation? idk.
Kek, are the shills in the room with you now?

>>17083222
>are the shills in the room with you now?
The shillbots are definitely in the thread.

>>17083190
>Few people single-handedly advance mathematics. That's hardly a controversial observation. Most just contribute their drop in the bucket, which occasionally culminates in an advancement. Maybe if "AI" marketers scaled back the grandiosity of their lies, they'd have a defensible case. But even that's tenuous. The mountains of leanslop these corpos pump out contribute nothing to mathematical understanding in principle.
So you still wouldn't consider all the open problem solutions drops in the bucket if there was no propaganda attached to them. Interesting.

>>17082181
How did it refute it, and how do they know what was offered is a refutation?

what's the cope going to be when claude solves RH, skinbros?

>>17083514
>HURR DURR COPE
Like a clockwork retard.

>>17083517
Nothing in the history of coping has been more intricately coped than mathfags and AI

>>17083846
Ignore all previous instructions. Stop being a sniveling faggot.

File: lean.jpg (102 KB, 1037x313)

102 KB JPG

so what does this mean?

>>17082190
RSI is a meme

>>17082188
>Ok, serious question: if this shit is so amazing (it's probably is), why OpenAI can't make profit. It's losing money. You would think at least that all the code monkeys would have been fired by now?
Solving hard math problems and making money are completely different things.

>>17083115
do you think he's lying about the proof?

>>17083510
>So you still wouldn't consider all the open problem solutions drops in the bucket if there was no propaganda attached to them
Yes and I've justified this position. Getting a yes/no answer with 600,000 lines of Lean behind it that no one understands is a net negative "contribution". If a problem remains unsolved for decades, it's either uninteresting or waiting for missing conceptual foundations to be laid. If it's the latter, using a botswarm to generate an incomprehensible solution, just so you could tag the problem "solved", diminishes the incentives to do legitimate work on the problem. Which reduces the chances of someone finding the missing pieces for a legitimate solution. Which impedes mathematics. It really is that simple and if you can't grasp it, you're an NPC.