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

It is truly a tricky area, I agree :)

Thought experiment: An array of GPT-3 agents trained on decade or century intervals of philosophical text/literature would have different ‘views’. Assuming the existence of mistakes, the post-enlightenment mistake is to assume the correct output is the latest GTP-3 agent.



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

Search: