Machine Learning

Cam aser: A lot of times i've researchers propose, agrthom that too computationally complex to run at the speed we need for production. Pu: I think that's probably also a an interesting thing to consider when you're doing the sort of prototype in getting things started. But vition, you talked about training a hundred thousand ways, right? Like, can you really spend 25 thousands dollars to decide on compute in the cloud or something side, if this is even going to work, right? Or could you do that in production and over the data you have, or something? Right? Pu, all thoughts on this.