Got any links to explanations of why fine tuning open models isn’t a productive ... | Hacker News

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

		bradfa 25 days ago \| parent \| context \| favorite \| on: 73% of AI startups are just prompt engineering Got any links to explanations of why fine tuning open models isn’t a productive solution? Besides renting the GPU time, what other downsides exist on today’s SOTA open models for doing this?

RC_ITR 23 days ago [–]

When the new pre-trained parameters come out in a new model generation, your old fine tuning doesn't apply to them.

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