High Signal Podcasts Evidence ledger
Method
Browse
← Back to evidence

Evidence receipt / evaluation

Published · transcript-backed

David Luan: evaluation

22 Mar 2024 Latent Space Why Google failed to make GPT-3 + why Multimodal Agents are the path to AGI — with David Luan of Adept

“One, I think that, frankly, a big part of it is that the public narrative is really forming around agents as being the most important thing. And I'm really glad that's happening because when we started the company in January 2022, everybody in the field knew about the agents thing from RL, but the general public had no conception of what it was.”

— David Luan

Source trail

Everything needed to verify it.

Speaker
David Luan
Attribution
Verified speaker
Claim type
evaluation
Recorded
22 Mar 2024
Publisher
Latent Space

Transcript context

…Why have you become more public? You know, if the whole push has... You're sold out, you're my enterprise, but you're also clearly putting effort towards being more open or releasing more things. I think we just flipped over that way fairly recently. That's a good question. I think it actually boils down to two things. One, I think that, frankly, a big part of it is that the public narrative is really forming around agents as being the most important thing. And I'm really glad that's happening because when we started the company in January 2022, everybody in the field knew about the agents thing from RL, but the general public had no conception of what it was. They were still hanging their narrative hat on the tree of everything's a chatbot. And so I think now one of the things that I really care about is that when people think agent, they actually think the right thing. All sorts of different things are being called agents. Chatbots are being called agents. Things that make a function call are being called agents. To me, an agent is something that you can give a goal and get an end step workflow done correctly in the minimum number of steps. And so that's a big part of why. And I think the other part is because I think it's always good for people to be more aware of Redept as they think about what the next thing they want to do in their careers. The field is quickly pivoting in a world where foundation models are looking more and more commodity. And I think a huge amount of gain is going to happen from how do you use foundation models as the well-learned behavioral cloner to go solve agents. And I think people who want to do agents research should really come to Redept. When you say agents have become more part of the public narrative, are there specific things that you point to? I'll name a few. Bill Gates in his blog post mentioning that agents are the future. I'm the guy who made OSes, and I think agents are the next thing. So Bill Gates, I'll call that out. And then maybe Sam Altman also saying that agents are the future for open AI.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.

Search evidence