Then why don't they hard-code the interface to the model to pretend you included... | Hacker News

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

		zahlman on Jan 4, 2025 \| parent \| context \| favorite \| on: Can LLMs write better code if you keep asking them... Then why don't they hard-code the interface to the model to pretend you included that in the prompt?

Phlogistique on Jan 4, 2025 [–]

...they do. It's called "o1".

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