Mistral CEO @arthurmensch says he doesn't subscribe to doomsday scenario's and doesn't see any meaningful paths in which AI escapes out of control in a way we can't just turn off the computers it runs on: "It's pretty clear that they are getting good enough to escape from non-hardened sandboxes. It's going to be a friction for adoption in the enterprise that autonomous agents have a tendency to do pretty surprising stuff". "Doomsday scenarios, I don't subscribe to them. I don't see any…