Skip to content

The dirty secret: LLMs run on other people's content

that's the dirty secret of all of this is that it none of these LLM is get trained without content. And where does the content come from?

Episode 169: Wifi & Web Apps (episode page, opens in a new tab) March 1, 2024, 2:47 p.m. CT 00:24:32 (play audio from this timestamp, opens in a new tab)

Speaker-labeled

Summary

As LLM aggregators start rewording and republishing creators' content for nothing in return, Dave points out the dirty secret: none of these models can be trained without content, and that content comes from creators.