Summary in Plain Language
Recently, OpenAI made a controversial move that caught everyone’s attention: CEO Sam Altman repeatedly called on the entire AI industry to slow down the pace of advanced research and development and to first establish unified safety standards. However, shortly after, OpenAI announced a major achievement—using 10,000 AI agents to work together and claiming to have solved one of the seven Millennium Mathematics Problems, the Navier-Stokes equation, in less than four days. They used a next-generation, undisclosed model that is even more powerful than the upcoming GPT-6. Yet, this achievement has not yet been officially recognized by the mathematics community and has sparked debates about priority and data theft. Essentially, OpenAI is in a desperate situation before its IPO, trying to slow down its competitors while simultaneously showing off its capabilities to boost its valuation.
---
Detailed Analysis
1. How credible is the claim that AI has solved a century-old math problem?
First, let’s debunk the hype around this “major achievement.” It’s far from being something that could be recorded in the history of mathematics. The seven Millennium Mathematics Problems were identified by the Clay Mathematics Institute in 2000, with a $1 million reward for solving any one of them. Over the past 20 years, only the Poincaré Conjecture has been formally proven, a process that took decades and involved cross-validation by the global academic community. OpenAI’s claim doesn’t even meet the basic requirements: according to the Clay Institute, any proposed solution must first be published in a legitimate academic journal and then wait for two years without any errors being found by mathematicians worldwide before it can be considered. OpenAI has not even gone through this formal process; they have merely shared a draft of their solution with the public.
What OpenAI is showcasing is not the mathematical result itself but a new approach to productivity. Previously, AI worked independently; you would ask a question, and it would provide an answer. This time, 10,000 AI agents (equivalent to 10,000 researchers) worked simultaneously, exchanging ideas, correcting mistakes immediately, and generating 2.7 million ideas and 130 billion words of content in just 88 hours—compressing the work that would normally take a human team several years into just four days. This ability of AI to collaborate on top-level research is what OpenAI wants to demonstrate to the capital market: their AI is no longer just a tool for chatting or writing copy; it’s a new form of productivity that can be used as a research team.
2. Academic controversy: The old rules no longer apply in the AI era
What was supposed to be a technological showcase has turned into an academic debate, highlighting the lack of clear guidelines in the field. Mathematicians at New York University and researchers at Anthropic have been working on the Navier-Stokes problem for a long time, with their results not yet published. OpenAI announced their solution without waiting for this process and even tried to exclude one of Anthropic’s researchers from the credit. OpenAI’s defense was rather dishonest: they claimed they didn’t steal any of their unpublished work, but it’s impossible to say with certainty that they didn’t use any of their research, given that both teams used ChatGPT in their research. This raises a significant question: if an AI company uses anonymous data from users’ interactions to complete a research project, does the result belong to the user or the AI company?
3. Altman’s “industry slowdown” advice was actually a strategic move
Many wondered why Altman kept urging the industry to slow down. Now it’s clear what he was doing: as the leading player in the AI industry, he wanted to slow down competitors while establishing safety standards that would be tailored to his own technology. By doing so, smaller companies and emerging competitors would struggle to catch up. The irony is that OpenAI’s actions show a double standard: while they called for a slowdown, they simultaneously showed off their advanced technology and announced plans to create AI researchers that don’t require human supervision by 2028.
4. OpenAI’s urgency: A race to go public
OpenAI’s bizarre behavior is a result of the intense competition with Anthropic for a spot on the stock market. Previously, OpenAI was seen as the undisputed leader, but recent data shows that Anthropic is catching up rapidly. Anthropic’s annual revenue (an estimated annual figure based on recent months’ earnings) has reached $65 billion, while OpenAI’s is only just over $25 billion. More American companies are using Anthropic’s services. With Anthropic about to go public and potentially attracting billions in investment, OpenAI needs to prove to investors that its technology is superior to avoid losing its leading position. By showcasing its next-generation model and its ability to solve complex problems, OpenAI aims to convince investors that its growth potential is much greater.
5. The industry’s new reality
This incident highlights a shift in the AI landscape. In the past, the competition focused on who could chat more intelligently, generate better images, or create better PPTs. Now, the competition has escalated to who can use AI for top-level research. The company that develops AI systems capable of collaborating on research will gain a monopoly in fields like drug development, chip design, and nuclear fusion. However, there are no clear rules regarding the ownership of research results or the protection of users’ privacy. The next few years will see leading AI companies pushing for rules that benefit them, determining their market share in trillions of dollars.