OpenAI is solving math... and nobody's happy about it

Crazy how many people just buy the stealing story hook line and sinker just because it fits their pre-conceived narrative. Read the details. The probability that the model used any of those previous chats in a meaningful way is vanishingly small.

“… then as if they just gotten vaccinated…”<br><br>🏴‍☠️

Yeah that&#39;s a big red flag for researchers.

Dude’s surprisingly uploading videos quite often than before he being sold off to a private equity, probably they are pushing him for more views or exposure

First. I stoned math

Damn I already was pretty hesitant to use chatgpt for anything personall. <br>I guess I will not use it at all in the future

QUESTION :<br>Can AI Models Solve Googological Problems, like Ordinals/Cardinals or Hyperoperations (Beyond Tetration) in Googology?

Typical mathematicians trying to gate keep knowledge

I just realized horse tinder isn’t such a bad idea after all, recently I found out they built a motel for horses where the owners take their horses so they can breed with other horses. maybe horse tinder could be the app to set up the meeting 😂

Ai all about stealing other people&#39;s working and pretend to be intelligent

Hard to believe no one would have thought to record that phone call.

Call me crazy but I don&#39;t really trust OpenAI about denying what they say they didn&#39;t say...

Can&#39;t believe that the Fireship video seems more accurate than the Two Minute Papers one, but here we are

I like how Navier and Stokes is the exact same guy.

AI can cure cancer if you paste the solution in the prompt input

OpenAI did the best advertisement for local models.

&quot;Open AI said...&quot;

Buckmaster

Imagine spending 20million$$ to solve a 1 million$$ bounty XD. What a clown show that didn&#39;t even work

Dude Buckmaster’s convo with the openai guy almost feels like a cliche movie villain trope. It’s scary how unapologetically greedy they are

Wait, isn&#39;t the point of these things that we would like to see proof? Why would mathematicians stop sharing data? That makes no sense, if these models can actually do this the incentive is to share more data so we get more proofs.<br><br>Or maybe... it&#39;s not about the proof but rather about the clout?

already on railway. great stuff

any updates on cancer cure yet?

Scam Altman

I&#39;ll wait for the 2033 fireship video

And the moral of the story: Always use local models for serious stuff!!!

The sponsor of your mom

The lesson here is local llms

I&#39;m Stoked by💲 Master in Navious ways...

&quot;Well, you can&#39;t prove it stole it! Guess you&#39;ll never know!&quot;

Why make a summary like this if you dont go all the way in order to actually estimate realistically if something relevant for navier stokes was &quot;stolen&quot; or not, looking at the forced and unforced euler / actual navier stokes pdf?

Hmmmm.... but when is AI gonna tackle the robot-waifu solution?

Okay Claude spin up 30,000 agents and let me know if P = NP, be careful and make no mistakes

So ai slop is happen in maths too

Innocent until proven guilty, no? Why would the scientists use chatgpt/claude, if it weren&#39;t useful? These are massive <b>large</b> language models, having a few text passages of supposed helping work isn&#39;t going to solve anything. The 10.000 agents solved it, unless the proof itself falls apart.

<a href="https://www.youtube.com/watch?v=aspmNhKAFMc&amp;t=129">2:09</a> that looks like an amazing friend to have.

Ok

Ada Lovelace&#39;s father = Lord Byron &quot;There is pleasure in the pathless woods, there is rapture in the lonely shore, there is society where none intrudes, by the deep sea, and music in its roar&quot;

Navier-Stokes is dead! Long live Navier-Stokes!!!

So company based on a theft of IP steals again and plays dumb again? What a surprise... /s

😂😂😂 they stole it, as expected of AI

Could OpenAI have fed the research to the agents? Sure. But I think it’s highly unlikely that the models were trained on their user data, considering that the knowledge cutoff for ChatGPT 5.6 Sol is February 16, 2026.

It didn’t solved it. It just stole it:)

They haven&#39;t even released the artificial movie yet and they might have the plot for the sequel 😆

Its wild that when AI does it, it is theft, but when ALL RESEARCHERS EVER do it, its fineeeee

Okay so Bubeck&#39;s post is the most reasonable. Buckmaster either intentionally or unintentionally misconstrued what was said. It is an extension of their work on Euler, but as a proof for Navier-Stokes it should stand on its own. It&#39;s a conventional expectation that Bubeck would be apprehensive to include Levent as an author on OpenAI&#39;s paper. He&#39;s trying to find compromise whereas Buckmaster is drumming up drama. These agents are making unprecedented headway in virtually all fields of research. I personally think it&#39;s moronic to forsake that for one&#39;s own pride.

No, OpenAI didn&#39;t solved.

Not to pooh-pooh on professionals, but what&#39;s more important; solving challenging problems or getting credit? Ideally we get both but if I had to pick...

Scam altman is real 😂

So, who is the one betraying us here?

I was waiting for this video

So basically they spent $20 million just to &quot;support&quot; what a <b>human</b> had already proved? And people are going crazy about it?

This raises massive questions about the future of open research. If AI can race to &#39;solve&#39; problems based on rumors of human progress, the incentives for collaborative science might disappear entirely. Really worrying trend.

well we thought chess drama was entertaining until math drama said &quot;hold my beer&quot; 😅😂

If you have any info worth exploiting Scam Altman is watching

Im surprised there&#39;s no vid from fireship about shoppify acquiring tailwind

Neatly summarises how AI companies work<br><br>Use 20x the resources <br>Plagiarize public / user data<br>Make a headline for something that didn&#39;t happen<br><br><br>God forbid anyone shares a cure to cancer in public domain <br>These mfs will gladly gobble that shit up and cry if anyone else tries to use it

The <b>you-ler</b> equations 🫥

Scam Altman strikes again

I think you need a new format. <br><br>By the time your videos came out I already seen 20 videos on the same topic.

Can he solve my kid&#39;s 2nd grade math?

The paper is really about proving a very specific Navier Stokes case in which a carefully constructed flow develops a finite time singularity, rather than solving every possible Navier Stokes flow. The important achievement is that the mathematical argument was also formalized in Lean, allowing the logical steps of the formalized result to be checked by a computer rather than relying only on human inspection. This gives researchers a rigorously verified piece of information about where the equations can fail and what kinds of behavior are mathematically possible. For CFD, results like this can eventually be turned into mathematical bounds, stability limits, validity conditions, and failure criteria that can be built directly into simulation software. The larger significance is that AI can now help discover and formally verify many such mathematical results, allowing future CFD programs to use proven bounds as computational guardrails instead of relying only on numerical approximations and empirical rules.<br><br>This prototype shows how future CFD software could use mathematical bounds to concentrate computation where the flow becomes most intense instead of using the same resolution everywhere. It calculates local vorticity (\omega=|\nabla\times U|), determines the volume weighted mean (\mu) and standard deviation (\sigma), and defines (B_{\rm low}=\mu+\sigma) and (B_{\rm high}=\mu+2\sigma), with the refinement indicator (I_i=\max\left(0,\min\left(1,\frac{|\omega_i|-B_{\rm low}}{B_{\rm high}-B_{\rm low}}\right)\right)). Cells with high (I_i) are then refined automatically while lower intensity regions remain coarser, reducing unnecessary memory and computation while preserving resolution where it matters. As AI discovers more rigorous bounds for velocity, gradients, energy, turbulence, stability, and error, future CFD programs such as OpenFOAM could turn those mathematical results directly into adaptive algorithms and solver safeguards that automatically decide where additional computation is needed.

OpenAI solves 1 million dollar problem by spending 20 million dollars on compute :)

The entire AI industry is based on stolen work.

The number of views is directly correlated to how many times per minute you change the thumbnail.

His voice sound different

Also Anthropic is almost opening IPO stocks

OpenAI stole the researcher’s partial solution and front ran him.

Why waste the first 10% of the video to explain what this video is about? Every one here is already here to watch this video.

Scam Altman doing scam altman stuff, what&#39;s new?

Gases are fluids. Saying &quot;fluids and gases&quot; is redundant.

Altman and OpenAI are the modern version of Edison and General Electric. Steal, sell, lawyer up, and win. Yikes.

&quot;Euler&quot; is &quot;oiler,&quot; not &quot;youler&quot; FYI

he is definitely an alpöge, but i dont see the buck master in the other.

Why use chagot products and expect privacy or them not training on your data

Astra, play Celery man

Alfa Gay such a cool name!

first 0.4 second was so funny .

OpenAI being unscrupulous? I&#39;m shocked!

Is coding largely solved or math largely solved?

Didn&#39;t the girl from gifted movie already solved this problem?

Am I insane or this EXACTLY what happened with anthropic and the Riemann hypothesis?

Wake me up when a frontier model writes a real proof for P = NP or the Beal Conjecture.

Me publishing first after seeing something cool

Well then... Brain tickled

You&#39;re telling me that Ada Lovelace would have a massive Umamusume fan!?

So OpenAI bad?

Spends 20Million to solve a problem and only get 1 million in return.

This here is going to explode in finite time😂

it&#39;s funny that the guy&#39;s name is claude-louis

no &quot;specific&quot; user data was accessed. ummm hmmmm.

The fact that OpenAI didn’t even claim the million dollars is the biggest flex

&quot;You tride&quot; haha

<a href="https://www.youtube.com/watch?v=aspmNhKAFMc&amp;t=51">0:51</a> <br><br>Of course his name was Claude

@<a href="https://www.youtube.com/watch?v=aspmNhKAFMc&amp;t=156">2:36</a> That is one way to pronounce Euler

Why am i first?

Scam Altman living up to his name once again

BRO, <a href="https://www.youtube.com/watch?v=aspmNhKAFMc&amp;t=324">5:24</a> broke me down 🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣🤣😂😂😂😂😂😂

Astra, solve Navier-Stokes. Show your work. &lt;steering&gt;…on second thought. 🙄

I&#39;d laugh if an LLM solves the Collatz conjecture.