LinkedIn Founder Reid Hoffman sees bright AI future and hopes his tech peers are right about Trump
(AP Illustration / Jenni Sohn) · Associated Press Finance · ASSOCIATED PRESS

LinkedIn founder Reid Hoffman has been immersed in Silicon Valley since his August 1967 birth in Palo Alto, California, in the shadow of Stanford University, where he and fellow technology luminary Peter Thiel became friends as college students during the 1980s.

They went on to start PayPal during the late 1990s while working alongside a coterie of other bright-eyed entrepreneurs who went on to even bigger things, just as Hoffman did. That group — dubbed the “PayPal Mafia” — included Tesla CEO Elon Musk, Yelp CEO Jeremy Stoppelman, and YouTube co-founders Chad Hurley and Steve Chen.

Now worth an estimated $2.6 billion, Hoffman has been at the forefront of the artificial intelligence craze while investing in trailblazing startups such as ChatGPT maker OpenAI and Inflection. Unlike other prominent technologists who are worried about AI destroying humanity, Hoffman has co-written a new book called “Superagency” that makes an optimistic case for AI. He recently talked with The Associated Press.

Q: Are humans creating something smarter than themselves with AI?

A: You already have some version of superintelligence right now. Obviously, we will be increasing that superintelligence, so we need to be thinking about which ways it's going to be making great additions and which ways it's going to make things more challenging.

Humans are generally bad at transitions, and that is part of the reason for writing books like “Superagency,” to try to say there are some positive design principles, some positive outcomes that we should be trying to work our way toward.

Q: Do you think there is a chance that these superintelligent AI agents eventually may seem omniscient?

A: AI agents are pretty useful for coaching and advising. Like if you were to say, “Hey, how should I approach this thorny problem?” An AI agent can be a pretty good advice-giving coach. And so you can imagine it being a useful tutor across a wide variety of subjects, including even spiritual or counseling or mental ones. And I think that could be positive for the evolution of humanity, having something talking to us that is infinitely patient and is trying to help us get to our better selves. So, yes, you could have a “Consult the Oracle” situation with AI agents.

Q: What's the risk of some people embracing AI as a bosom buddy and becoming more socially isolated?

A: Some of that will happen for sure, the same way some people isolated themselves with TV or the internet. Part of this will be about the design focus. If you go to Inflection's AI agent, Pi, and say, “Hey you’re my best friend,” Pi, will say, “No, I am your AI companion, let’s talk about your friendships. Have you seen any of your friends recently or want to see them?” We want AI agents being softly and gently encouraging, to be tied into the human circumstance, which I think is what most people really want.