needing to kind of be a scientist being like, "Oh, I got a good idea. Let's try that." Okay, 5 weeks later, "Okay, I got another good idea. Let's try that." And that all gets reduced to silicon. >> And are you using your own model? Are you training your own AI, or you using some of the publicly available frontier models? >> In my world, we use both. So, we take our own data and then we build our own predictive models from it. And then separately, we use off-the-shelf life sciences models, and then we also use some of the general models. So, it's a great combination. And to your questions and talking about AI earlier, oh my god, the stuff we can do is unbelievable and it's mind-blowing. And
Are you training your own AI, or you using some of the publicly available frontier models?