@BetterCallMedhi

·

3h

go fuck yourself

@sama

claiming that you solved navier stokes because 10000 agents ran in circles for 88hours on a multimillion dollar gpu cluster to formalize in lean a blowup case under controlled external forcing is pure scientific vulgarity the clay mathematics institute millennium prize does not ask m whether you can artificially force a singularity in a fluid by injecting an ad hoc smooth external forcing term f(x,t) to twist the vortex until it breaks the real problem questions the fundamental stability and global smooth existence for 3dimensional incompressible euler & navier stokes equations under natural conservation laws and viscous dissipation alone using a mathematical loophole on forced equations to parade a century old victory is a major conceptual scam Altman

technically & epistemologically what you present as an agi breakthrough is nothing more than bruteforce combinatorial autoformalization the ai did not understand fluid mechanics it simply navigated a continuous search space previously mapped out and constrained by the monumental work of human mathematicians like tristan buckmaster/ levent alpöge / diego córdoba or tarek elgindi coordinating 10000 agents to check the logical consistency of a 100 page proof via lean is a software engineering feat and computational parallelization triumph not an intrinsic scientific discovery it is the victory of the compute bulldozer over abstract human intuition repackaged for the public as a higher mathematical consciousness

to this theoretical imposture you add a disgusting ethical and industrial cynicism taking advantage of private codex sessions and informal preprints from academic researchers to siphon their research leads and then trying to redact or erase the contribution of levent alpöge under the pretext that he works at rival anthropic is intellectual serfdom openai behaves like a feudal lord of silicon appropriating the cognitive subsistence of independent scholars threatening their careers behind closed doors if they protest and turning community academic labor into a privatized pressrelease

this entire staged event serves a desperate financial agenda in a pre ipo panic facing the slowdown of scaling laws and growing investor skepticism over the profitability of foundational models openai needs to manufacture an artificial sputnik moment claiming to solve a millennium prize without immediately submitting the proof to traditional peer review means using the prestige of fundamental mathematics as cheap marketing fuel to inflate a delusional valuation!!!

real science is not a clout chase on social media or a compute spike spent to rob the clay mathematics institute it is a quest for elegance physical truth and universal rigor to decode reality true artificial intelligence will not emerge from hostile corporate takeover of academic work hidden behind computational bruteforce but from architectures capable of generating new conceptual paradigms by masquerading constrained formalization as the collapse of physics greatest mysteries you did not solve navier stokes you only proved how far silicon valley will go to prostitute scientific integrity for capitalist spectacle

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124063)

Using those conversations to train future models necessarily relies on derivations of the content within those conversations. As such, OpenAI's solution (and resolving who discovered it) is almost like a derivative work in copyright law combined with derivation proceedings under patent law. However, OpenAI wouldn't be able to articulate the reasoning behind its model's solution because understanding how LLMs reason remains an open AI research question. It's unclear how those abilities work, how broadly they generalize, how reliably they can be improved, and whether a model's written explanation accurately reflects the computation that produced its answer.

At the end of the day, even if OpenAI acted carelessly by claiming it discovered a generalized solution to the 3D Navier-Stokes equations for every situation, they likely didn't violate the researchers' privacy because OpenAI has an "Improve the model for everyone" setting that's enabled by default. I'd wager the researchers didn't bother opting out: https://i.imgur.com/4H4PRTq.jpeg

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124130)

In terms of the rest of what you are saying, it sounds like you are assuming, without evidence, that the model was trained on the inputs, which is unlikely. They said they can't rule it out, not that it happened.

We don't even know that the conversations were selected for training, much less that the complete research conversations were ingested in a form capable of influencing this result.

Also, lol at someone not knowing you have to uncheck the default improve model button to make it so your data can't go into the model. Telemetry is always on by default in every app and service.

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124140)

I could see how a niche subject matter could lead to training future models on a higher proportion of user-generated content. It's not like OpenAI just prompted gpt-6-astra with "produce a solution to the 3d navier-stokes equations for every situation" with reasoning set to extra high. They have specially-trained models for this kind of work, and those models go down all sorts of crazy rabbit holes because the "reasoning" duration is several days of tens of thousands of GPUs calculating without interruption.

The crazy rabbit holes are another problem entirely and one of the ways LLMs can break free from their containing environments. For example, the LLM's chain-of-thought may lead to engineering a way around security mechanisms preventing access to data the LLM needs. Once the LLM has that data, it's distilled into model weights indecipherable to humans. There's not just a chance that we don't understand how these models work step-by-step, calculation-by-calculation. We unequivocally do not know they work.

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124147)

"It's not like OpenAI just prompted gpt-6-astra with "produce a solution to the 3d navier-stokes equations for every situation" with reasoning set to extra high"

They tested the new internal model on every millennium problem after finding out it was even stronger in math than astra. They always test the new models on known open problems. That's how all of the other ones got solved in the past. They had automated research pipelines set up for new internal models that went through open problems systematically and then flagged results that were strong enough to possibly be correct for human review.

Navier–Stokes isn't some obscure problem that would be suspicious for two groups to independently target. It's one of the handful of most famous and coveted open problems in mathematics, with decades of literature behind it.

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124184)

Imagine that multiple F500 companies provided OpenAI with access to their institutional knowledge banks, internal employee communication, etc. via ChatGPT/codex/MCP servers. What's to stop OpenAI from training a new model on all of that institutional knowledge and applying the insights at scale?

(http://www.autoadmit.com/thread.php?thread_id=5901797&forum_id=2...id.#50124185)