Matt Wolfe explains how AI reasoning models are changing disease diagnosis. These models outperform doctors in accuracy, and hospitals are starting to integrate them into real diagnostic work.
Based on the data that already AI like reasoning models are doing better than doctors on solving hard to diagnose diseases. So as of right now, and I think the numbers was 80% versus 30%. 80%. Like in terms of hard to diagnose diseases, that when you were testing a doctor, if the doctor could diagnose the disease, the doctors got it right 30% of the time. The AI got it right 80% of the time. And this is not the new models. This is the, the very first preview of O1. And from a lot of stuff I've seen, Zero1 Pro is probably in the ballpark of three times smarter than that model. So probably when the new data out, it's gonna be like, okay, it's not 80%. Like, it's 95% or 90% and the doctors are 30%. That's a huge difference. So this just shows, I guess that's one reason we started the podcast. Right. Like, I think most people don't realize, like, this is society changing stuff. Like, this should be where we restructure society, where, yeah, you still need doctors, but you have doctors who are highly relying on AI to help diagnose diseases.
AI News: Dots, GPT-6.1 Sol, Sonnet 5.5, Gemini 4, and everything you need to know
