A matematikusok szerint az OpenAI félreértelmezte a Navier-Stokes-problémára benyújtott 166 oldalas megoldásának módszertanát. A cég eredetileg azt állította, hogy a bizonyítás két spanyol matematikus megközelítését követte, a részleteket vizsgáló szakértők szerint azonban az alkalmazott eljárás csak a felszínen hasonlít erre. A Lean formális ellenőrzőnyelven írt anyagnak jelenleg is zajlik a szakmai felülvizsgálata.
A kutatást komoly adat- és módszertani viták övezik. Nyilvánosságra hozatal előtt kutatók azzal vádolták a céget, hogy mások nyilvános munkájára épített, amit a vállalkozás tagadott. Ráadásul az OpenAI kutatója, Noam Brown elismerte, hogy nem tudják pontosan, mekkora szerepe volt a végeredményben a feladatra ráeresztett, 10 000 modellpéldányból álló ágensrendszernek. A hatalmas erőforrás-igény miatt az ilyen méretű ágenshálózatok működését még nem tudták szisztematikusan tesztelni.
Az eredeti szöveg (The Decoder)
In early September, the company presented a 166-page potential solution to the Navier-Stokes problem. OpenAI claimed the proof followed the approach of Spanish mathematicians Diego Córdoba and Luis Martínez-Zoroa.
Several groups of experts now disagree, Scientific American reports. "When you go to the details, it is a different approach," says mathematician Gonzalo Cao-Labora of EPFL. The similarities are only surface-level, he says.
The potential proof is written in Lean, a formal verification language that checks every logical step but doesn't make it readable to humans on its own. The Clay Mathematics Institute (CMI) has been reviewing it since its release.Ad
Before publication, NYU mathematician Tristan Buckmaster accused OpenAI of building on work that he and Anthropic's Levent Alpöge had done with publicly available models. OpenAI denied the claim, saying it never used Buckmaster's data. A definitive answer is probably impossible because language models are black boxes, according to Camillo De Lellis of the Institute for Advanced Study.Ad
In early October, OpenAI dumped roughly 400 more results that the math community can barely keep up with. Publishing them as a loose collection on GitHub angered many mathematicians, who felt it cheapened their field.
There's another thing OpenAI doesn't seem to understand about its own work. The company threw a lot of compute at the effort, deploying about 10,000 model instances for roughly 88 hours and 130 billion tokens to develop the potential Navier-Stokes solution. According to OpenAI researcher Noam Brown, nobody actually knows what that massive agent swarm contributed to the result.Ad
Behavior at that scale hasn't been studied because running systematic comparison tests with thousands of agents costs too much. No one has tested how long a single agent would take to solve the same problem, and it's nearly impossible to tell what concrete advantage 10,000 agents offer over 1,000.
Brown puts less than 10 percent of the success down to the multi-agent setup. These systems are new and flashy, he says, so they get a disproportionate share of the credit. What really matters is the base model itself, which can grind through tasks over long periods and run in parallel.Ad
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
