Shawn Wang: uncertainty

Transcript context

…Yeah. I'm really curious how ramped does the attribution to the model? If it's API, because I think it's all credit card spin. . Well, but it's all, the credit card doesn't say maybe. Maybe the, maybe when they do expenses, they upload the PDF, but yeah, the, the German I think makes sense. I think that was one of my main 2024 takeaways that like. The best small model companies are the large labs, which is not something I would have thought that the open source kind of like long tail would be like the small model. Yeah, different sizes of small models we're talking about here, right? Like so small model here for Gemini is AB, [00:13:00] right? Uh, mini. We don't know what the small model size is, but yeah, it's probably in the double digits or maybe single digits, but probably double digits. The open source community has kind of focused on the one to three B size. Mm-hmm . Yeah. Maybe zero, maybe 0.5 B uh, that's moon dream and that is small for you then, then that's great. It makes sense that we, we have a range for small now, which is like, may, maybe one to five B. Yeah. I'll even put that at, at, at the high end. And so this includes Gemma from Gemini as well. But also includes the Apple Foundation models, which I think Apple Foundation is 3B. Yeah. No, that's great. I mean, I think in the start small just meant cheap. I think today small is actually a more nuanced discussion, you know, that people weren't really having before.…

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