AI-rewritten: This is a summary of an article from Tom’s Hardware, rewritten by AI (Qwen, running locally) to make it easier to read. The facts come from the original article – read it for the full story.

---

OpenAI recently released over 722 manuscripts containing mathematical solutions generated by an unnamed internal frontier model. While analysts are working to verify these results, mathematicians warn that validating them could take years. The field faces concerns because the AI often jumps directly from a prompt to a solution without showing the intermediate steps required for human understanding. Experts argue this lack of transparency prevents others from building on the work and questions the legitimacy of the research itself.

The situation has raised serious fears regarding cryptography and national security. Some experts suggest that if AI can solve complex mathematical problems, it might eventually defeat current encryption methods. This could put trillions of dollars in cryptocurrency at risk. However, not everyone agrees with this prediction. Cryptographer Adam Back called these claims nonsense, while Ethereum creator Vitalik Buterin urged the community to remain calm and avoid panicking over wallet security.

Mathematicians are deeply concerned about the impact on their careers and the collaborative nature of the field. Professor Tristan Buckmaster noted that OpenAI’s releases have upended early-career researchers’ work, creating an atmosphere of anxiety. Professor Bryna Kra added that the rapid pace of AI output has made mathematicians more secretive, hindering collaboration. Terence Tao argued that the field may need to value community building and holistic progress over raw solutions. With rumors of further releases from other labs, morale remains low as researchers grapple with the uncertainty of their future work.