“If there is what they call their p-doom, their probability that it will doom us all, and some of them on the Builders podcast say it's 10%, I go, wait a minute, if it's 10%, and if it's 1%, if it's a non-trivial number, shouldn't we be pausing? I think all of them in some ways would like to pause. But it's such a competitive environment that Anthropic's not going to pause, because maybe OpenAI steps in and says, we're not going to pause. Or they say we're in a race with China.”
EconTalk · In-Depth Interviews · November 2026
Context
Roberts asked whether the AI builders Isaacson interviewed feel real fear about what they are making. Isaacson describes it as a prisoner's dilemma: each lab feels it can't slow down unless the others do.
Transcript
…Gates, you know, Steve Jobs, when Steve went to college, he took calligraphy and dance and music and poetry. And Bill Gates took applied math and mainly math and science courses. He said, if. Steve, if Bill Gates had taken some more humanities courses, he might have understood how beauty matters.
…talking about Steve Jobs, you mentioned that technology married with liberal arts, married with the humanities. It's that that yields us the result that makes our hearts sing. Now, these folks don't seem to be doing a lot of singing. Do you sense genuine fear about the world they're creating?
I think that they do believe that there is a chance that it will go totally awry. And in the Builders podcast, whether it's Elon Musk or Dario Amadei or Mustafa Saleman, I asked the basic question: why are you building this technology? If they say it creates enormous, I say, but does it help humans? Does it help humanity? And if there is what they call their p-doom, their probability that it will doom us all, and that p-doom is some of them on the builders podcast say it's 10%. I go, wait a minute, if it's 10%, and if it's 1%, if it's a non-trivial number, shouldn't we be pausing? I think all of them in some ways would like to pause. All of them think there should be some guardrails, but it's such a competitive environment that Anthropic's not going to pause and put too many guardrails because when it does, and when it tells the Pentagon it's going to put these guardrails in pause, maybe OpenAI steps in and say, we're not going to pause. Or they say we're in a race with China. So sometimes you get into a prisoner's dilemma or some place where you know you should be more cautious, but the competitive race makes it difficult to do so.
You know, I don't know all the people on your podcast. I know a few of them. And in general, my experience of Silicon Valley, which I spent many summers at Stanford, it's an extraordinary place. There's so many brilliant people putting dents in the universe, but their knowledge tends to be quite…
think that's true. I think, you know, if you listen to Dario Amadei and Daniela, I interviewed them together, his sister. You know, they understand philosophy and the arts, especially Daniela. They've hired Amanda Aska, not hired, but on their team is Amanda Askell, a great philosopher from Oxford…
I hope you're right. There's certainly some really broadly read people and broadly educated, but in my experience, there's a lot of narrow ones and they're on a mission, which I get. Some of it's competition. Some of it is their own desire, understandably, to, as I say, to put a dent in the…
More from EconTalk
