Can We Make This Work for Who?

Jake: Can we fine tune a model to basically be able to code in who better answer questions about whom which would be amazing for people who are learning it. I think the amount of work coming out and I'm familiar GPT for all is really impressive. The fact that people are getting these kind of performance gains just by doing technique based stuff and not by increasing the size of the models implies that they're doing that because they think that's what open AI did to get GPT four. That also has really big implications for the future of what's a college GPT five.

Play episode from 38:41

Transcript

The AI-powered Podcast Player

Save insights by tapping your headphones, chat with episodes, discover the best highlights - and more!

Get the app