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

Definitely true for GPT-3, but after using GPT-4, I am not so sure. I am sure there are cases where a specialized NLP model can beat GPT-4, but I think they might be in the minority now. Whether I tried summarization, data extraction, or code generation, GPT-4 just kinda worked.


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

Search: