Q and then we'll open up to people, just, uh, a handful of questions. Um, so machine learning is great at sifting through massive amounts of information, but it's not always great at being, uh, reliable, right? It works for most of the cases, but not all of the cases. Uh, how do you think about responsibility, I guess, in terms of not making the right recommendation or the right analysis?
A So I think that's, ah, that's a crucial, ah, element of what we're doing. So bear in mind, at stage one, and the app that we have outside replaces WebMD. So I look at it and say, wow, if people, sixty million Americans every month are willing to use something that doesn't work, well, if they use something that works pretty good, you know, it's so much better. And the reality in medicine, There's no absolute, ah, solutions. You don't know. Ah, you ask doctors. If you go and speak to ER physicians that have seen enough cases over the years, they'll always say horrible things happen. So there's no a hundred percent accuracy. But bear in mind, when you go to a doctor next to the Harvard or Stanford medical degree, it doesn't say 92% accuracy for the last 5000 people. They don't measure accuracy. They don't even know how to measure it because they don't have a closed loop of data. So I think you need to build these things right. If we sit here and say, okay, somebody might die and people will die. And therefore we shouldn't be helpful to people. We're just doing ourselves a disservice around heart attack, stroke, diabetes, acute, uh, all these different things. I don't think medicine at the primary care certainly hasn't changed since our parents were kids.
AI assessment note: “If we sit here and say, okay, somebody might die and people will die.”