The Future of AI

I think people are still trying to absorb what's possible. You want some determinism out of it. QAing this stuff is like completely insane. What do you do with something that 97% of the time does this amazing thing? And then, you know, one out of 100 times does something really weird. We're just not used to thinking about things quite that way. I don't know. Is the average MLOS person equipped to monitor agents? IDon't know, maybe.