Recently, rumors about xAI’s large language model Grok3 proving the Riemann Hypothesis caused an uproar in the AI circle. xAI engineer Hieu Pham broke the news on social media that Grok3 successfully proved this mathematical problem and therefore suspended training. As soon as the news came out, it immediately aroused heated discussion. However, Pham soon clarified that this was just a joke, an oolong caused by netizens' revelations about a "catastrophic incident" during Grok3's training. The editor of Downcodes will take you to review this "farce" and discuss the progress of AI in overcoming mathematical problems.
A few hours later, Pham revealed the answer in another post: It was just a joke. This "own goal" originated from the revelation of netizen Andrew Curran, who claimed that Grok3 encountered a "catastrophic event" during training.
Faced with increasingly outrageous rumors, xAI co-creator Greg Yang couldn't help but post a sarcastic post, "Yes, yes, yes, Grok3 started attacking office security after training." Another researcher, Heinrich Kuttler, also humorously said He said, "The situation was very bad! We later replaced all the bad weights with nan (Not a Number, non-number) and then restored it." Upon seeing this, netizens also joined the ranks of making memes.
Although this "farce" ended in a joke, it also triggered people's thinking about AI's mathematical capabilities.
So, how far is AI from solving millennium mathematical problems such as the Riemann Hypothesis?
We can get a glimpse of the performance of AlphaProof, an AI mathematical proof tool developed by Google's DeepMind team. AlphaProof successfully solved three questions in the 2024 International Mathematics Olympiad (IMO). The sixth question is known as the "ultimate boss" and is extremely difficult. AlphaProof demonstrated strong logical reasoning and creative thinking during the problem-solving process. For example, in the second question, it cleverly chose to consider the number ab+1 to construct the proof. This strategy is consistent with human problem-solving ideas. And together.
Although AlphaProof has achieved impressive results, AI still has a long way to go to overcome top mathematical problems such as the Riemann Hypothesis. The Riemann Hypothesis has a history of 165 years since it was proposed in 1859. Countless mathematicians have devoted their efforts to it, but they have never been able to completely prove it.
To prove the Riemann Hypothesis, AI needs to have powerful computing power and deep reasoning capabilities. Currently, AI can find provable theorems by exhaustively searching all possible proofs, but this requires astronomical amounts of computing resources. In addition, AI also needs to have the ability to understand and apply existing mathematical tools in order to play a greater role in mathematical research.
Some AI experts predict that by the end of 2026, AI will become a "supermathematician" capable of solving difficult problems such as the Riemann Hypothesis. Musk has also promised that Grok3, trained with 200,000 H100 units, will be released by the end of the year and will bring amazing performance.
Let us wait and see whether AI can achieve breakthroughs in the field of mathematics in the future.
This "own incident" regarding Grok3's proof of the Riemann Hypothesis not only demonstrates the rapid development of AI technology, but also reminds us that we should remain cautious and rational in our assessment of AI capabilities. AI has broad application prospects in the field of mathematics, but overcoming century-old problems such as the Riemann Hypothesis still requires a long and arduous exploration.