Do you believe OpenAI's AI really solved hundreds of open math problems, before experts had a chance to peer-review the results?

OpenAI released 722 manuscripts from a yet-unrevealed frontier model, grouped into 372 result families — the company says they solve hundreds of open math problems. Researchers warn this raises serious research-ethics and academic concerns, since the work hasn't passed peer review yet.

• Yes, I trust this is a real breakthrough
• Maybe true, but I won't take it at face value without independent validation
• I'm skeptical until peer-reviewed publication confirms it
• I see it more as a marketing stunt than real science