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

Because things like GPT have a limited contextual memory. When you exceed that token limit it starts forgetting things, which becomes problematic on long pieces of code.

Just copying the latest code every once in a while (as long as it isn't too long) and putting it in a new window commonly works around issues like that.



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

Search: