Gurley predicts restricting open source AI in US while leaving it available elsewhere will shift innovation abroad.
“if you pull up and say there's no open source here and you leave open source there, guess where it's going to be the most innovative, where where you're going to have the most entrepreneurs, where you're going to have the most innovation. It's gonna be over there,”
Gurley argues AI crimes are already illegal under existing laws, questioning need for new regulations.
“If I defraud you with AI, it's still illegal as it would have been if I'd have done it with a phishing scam before AI. And so there why do we need new laws?”
Gurley argues biggest AI doomers are those with largest financial stakes in foundational models, not disinterested academics.
“the biggest doomers are the ones that have the biggest financial stakes and the biggest foundational model.”
Gurley questions why advocates believe Washington would be effective at regulating nascent AI technology.
“The thing I would ask these these people is why do they have confidence that Washington would be good at regulating something this nascent? Like, why would they be effective?”
Gurley observes the biggest AI doomers have the largest financial stakes in foundational models.
“It's remarkably interesting that the biggest doomers are the ones that have the biggest financial stakes and the biggest foundational model.”
Gurley says state AI laws stem from regret over missing social media regulation but create massive fragmentation risk.
“And and and I I think some of that comes from there. You run this massive risk of of trying to regulate a brand new technology at a state by state level.”
Gurley argues 50 different state AI rules will slow US companies competing globally against unregulated foreign rivals.
“Right now, a lot of people believe we're in this global competition to to see, you know, whose tech stack for AI is used on a global basis.”
Gurley argues Anthropic and OpenAI seek regulation to handicap startups and open source competitors.
“I think for certain reason Anthropic and OpenAI are begging for regulation is so that they can write rules that make it harder for startups, open source models to be competitive with them, which I think is very rational for them to do, but I don't think it's good for society.”
Gurley predicts US AI regulation could create a fence letting China serve the global market instead.
“I think what could happen with AI is we could build a fence around ourselves, and China could end up serving the rest of the world.”