Hacker News Reader

Hasn't gpt-3.5-turbo fine tuning been out for only a couple weeks?

How fast are people supposed to pick up these bleeding edge capabilities (like productized fine-tuning) when even the class of technology is so new?

Fine tuning of other "completions" type models has been available for a couple years

I get that but it's been off many organizations' radar completely until 3.5 became available.