Why AI needs physics to stop guessing

It is, yeah. Yeah, you mentioned, what was the phrase you said, in physics you trust and in AI... We believe. So, I mean, at what point do you stop believing and begin trusting or vice versa? Well, I think that subtlety is because AI is still probabilistic, right? Physics is deterministic, right? And because we need chips to have defined states, it has to be... deterministic by definition, right? That's why the nuance, but really it's evolving, right? As agents are becoming more and more embedded in the flows, you know, we rely more and more on the agents to really explore more options, close coverage faster. So it's an evolution, right? It started with reinforcement learning, and then it went to generative AI, and then it went to AI as an assistant. And now it's really becoming AI as part of the autonomous flows that are coming with agents. Part of it, which means engineers still have to babysit the information, obviously. Indeed, because they have the judgment and the context that AI is still learning, and they don't have the experience yet.