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

Oh, I think I see. You're imagining LLMs that learn from PO feedback as they go?


This can be as simple as giving them search over communication with a PO, and giving them a place to store information that's searchable.

How good they are at this is a different matter but the article claims it is impossible because they don't work on the code and build an understanding like people do and cannot gain that by just reading code.




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

Search: