“I don't think we know how to regulate AI safety. And we don't, in part because of the nature of the technology, this is not deterministic software code. This is probabilistic. Nobody knows exactly what comes out of these models… The expertise is nowhere near Washington. It really is in the labs.”

me explain where I was going. The reality is, I don't think we know how to regulate AI, AI safety. And we don't, in part because of the nature of the technology, this is not deterministic software code. This is probabilistic. Nobody knows exactly what comes out of these models. So it makes it very different as we think about regulating. But part of the challenge, too, is that the expertise is nowhere near Washington. It really is in the labs. And we have heard plenty of cries for more AI safety regulation. We have to engage the leading thinkers. Who are overwhelmingly in places like Greg's district in my district? They're not in Washington, D.C., to actually enable us to create much safer AI. And ultimately, I believe it's going to require us to engage the industry proactively with obviously the right liability rules and so forth to…