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

If the reasoning model was truly reasoning while the flash model was not then by definition shouldn’t it be better at knowing when to use the tool than the non-reasoning model? Otherwise it’s not really “smarter” as claimed, which seems to line up perfectly with the paper’s conclusion.


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

Search: