Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> maybe you have examples of historical breakthroughs in AI Alignment

OpenAI confessions is a good example of largely non-mathematical insight:

https://arxiv.org/abs/2512.08093

I don't know, I think it's good stuff. Would you agree?

> I think it's possible that some people are talking to ChatGPT on a buzzword level

I never said this is not happening. This definitely happens.

What I said is very different. I'm saying that you don't need to be a mathematician to have good insights into novel ways of improving AI alignment.

You definitely need good epistemic intuition though.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: