03: The Next Generation of LLMs with Jonathan Frankle of MosaicML
Replit AI Podcast
00:00
The Bias of Open Source Data
A lot of people have put a ton of work into these data sets. But it's really hard work to do and it's unglamorous work. So we've done incredible work so far, but a lot of our data sets are just old. I don't think my friends at a Luther would be mad at me if I said the pile is an old data set"
Transcript
Play full episode
