Elon Musk on Sam Altman and Anthropic
Walter Isaacson put the question to Elon Musk directly: would he ever team up with Sam Altman? The reply landed hard, "So deceptive, it's crazy." Then the conversation pivoted to Anthropic, and the temperature dropped a few degrees. Musk said he visited their people before renting them compute, running a vibe check to see if anyone tripped his evil detector. His conclusion? "Everyone I met at Anthropic was very nice and conscientious." Nobody came off shady or deceptive. Which is a whole arc, considering he'd previously branded the company "misanthropic and evil." Now the compute is flowing their way regardless. He also floated a bigger idea for the five frontier labs: weekly calls, the kind where everyone asks how to save humanity from itself. Why bother? Because standing up a regulatory agency drags on forever and the rulemaking crawls. His alternative has competitors kicking the tires on each other's models a week before launch, so if someone reports "oh, I was able to make a nuclear bomb with your AI," it gets patched before release. OpenAI, from the sound of it, isn't in the room for that one.
Transcript (en)
Well, I guess we are providing them with the compute, which they really needed. But before I ran into the compute, I went and met with a bunch of people at Anthropic to sort of, I don't know, see if anyone set off my evil detector or anything like that. Not that my evil detector is perfect, of course, but you know, might as well just try, you know. And everyone I met at Anthropic was very nice and conscientious. So So I didn't meet anyone who seemed shady or deceptive or anything. Could you imagine the five frontier models, yours and the others, just getting together once a week and saying, all right, how are we going to save humanity from ourselves? No, in fact, I've recommended that we do that exact thing because there's a lot of talk of having a regulatory agency. But the thing is that regulatory agencies, it takes a long time to get one established. and their rulemaking is very slow. And I said, like, guys, what if we just have a call once a week, just like you're saying, and that maybe we, a week before we release a model, we give that, you know, we give access to our competitors to say, hey, guys, why don't you try out the new model, see if you see any problems with it. I think competitors would keep each other honest in that situation. You know, if they say, oh, it's, you know, I was able to make a nuclear bomb with your AI. It's like, oh, we better fix that before we release it, you know, type of thing. So, I don't know. I guess I'll talk to Anthropik about this. Could you talk to Sam about it, or is the poison too strong there? He's so deceptive. It's crazy.
