Steven Bartlett had heard enough resignation for one conversation. If his guest Konstantin Kisin genuinely believed that AI was an existential threat, and genuinely believed that on the current course of travel it was inevitable, then surely, Bartlett asked, "Shouldn't you be doing something about it as well?"
"I can't do anything about it," Kisin said.
Bartlett pressed: who can? "No one." Not Trump? "What's Trump going to do?" In Kisin's account the technology arrives whatever anyone in a studio decides: "It's going to happen one way or another, and we don't know what that's going to be like." And then the line the rest of the exchange circles around: "But there is no solution to that issue."
That disagreement — between private acceptance and collective influence — runs through the back half of the episode. Kisin, a writer and podcaster, argues that the only remaining choice is personal. Bartlett, whose own show is part of the public conversation about AI, keeps looking for the lever that has not yet been pulled.
"That's all that's left"
Kisin calls his answer spiritual, by which he means something fairly plain: "ultimately none of us can control the things that we can't control." So the question becomes how you live, knowing what you know.
He says he watches what people do rather than what they predict. His example is people who "catastrophize about the climate, but buy a big house on the seafront" — a gap, he argues, that tells you what they truly believe. He applies the same test to AI. If AI is going to kill everyone, then the best thing available is "to enjoy your life now, to enjoy time with your family, your kids," and to do things you think are meaningful.
Then he turns it on the host. "Are you going to stop AI, Steve? You're powerful and successful and influential." His own answer: "You're not going to stop it." He counts out the people in the room — "I don't think Steve is. I don't think I am" — and lands on what he thinks is left: to "live in a way that is beneficial to us, our families, the people around us, our community, our country. That's all that's left."
Bartlett's lever: the public conversation
Bartlett does not accept the framing. "I would think that we can have an impact, though," he says. His evidence is the way public sentiment has already turned: he describes Eric Schmidt being booed by 7,000 students at a university commencement speech, and calls it "a consequence of what a lot of us do" — people putting information into the world and holding the argument in public, something he says would once have been blocked by gatekeepers in the media.
From there he looks ahead to the 2028 election cycle and the issues that will be on the ballot. "I think AI is going to be a massive one."
Kisin thinks that is optimistic. Most people, he says, "don't get this conversation" — the panel talks about it constantly partly because they have all been to San Francisco and spoken to people there, and "the overwhelming majority of people don't have that experience." For most, he says, the deepest exposure to AI is having tried an early model two years ago, finding it "totally inaccurate," concluding "this is a bit shit," and getting on with their lives.
Bartlett tests the forecast against a specific scenario: within two years, unemployment rising particularly at the entry level of employment, while on the other side of the ledger Anthropic reaches a $10 trillion valuation — a hypothetical he puts to Kisin, not a company projection. People, he suggests, will make the connection.
Kisin doubts it. "They'll say, I'm unemployed because Trump is bad." Or because some other politician is bad. "That's how people think, unfortunately." In his version the anger arrives, but it lands on incumbents rather than on the technology or the companies building it.
Inside what Kisin calls the gold rush
Kisin's fatalism rests on what he says he has picked up from people he has talked to — not, he stresses, the very people at the top, though from what he hears they are similar. He describes "a messianic cult in Silicon Valley of people who believe that they're on the frontiers of being God," mostly young men "very, very seduced by the idea of being God as young men have always been." They are not exactly addicts, he says; it is a gold rush, "except it's not gold. It's power. It's destiny. It's the ultimate technological breakthrough. It's your name and history" — motives he calls "way more powerful than gold." His conclusion: "there is no way in hell they're going to stop."
Bartlett says he has heard similar things in private conversations, through friends who know these people, and that he thinks Kisin is right.
What follows is the part Kisin insists on: that those arguing for speed have very good arguments and deserve credit for them. He performs their case in the first person. I, with my frontier model, am going to solve cancer. Solve inequality. Make sure human beings never fight another war. Make sure no child ever dies in infancy. "How can you be against this? How can you say we must go slower?" Why be careful, he has them ask, "when I'm saving lives, billions of lives, humanity?" — saving it from nuclear disaster, from climate change, or simply going harder so that five years from now there is life on another planet and the climate question no longer bites. And China is coming, which in this telling is not a worry but a bonus: "The competition will make us all get there faster. Isn't that wonderful, Stephen?"
Why climate is different, in his account
Bartlett returns with the inconsistency he has been holding. Earlier in the episode, on climate change, Kisin had more optimism for humanity; he would not concede there was nothing to be done, even as the conversation turned on the argument that it was a bit late, that we were "shutting the door after the horse has bolted," and that humans tend to unleash things and find out the consequences afterwards. When the economist Steve Keen said we were five seconds from the brick wall, Bartlett recalls, Kisin said no.
Kisin's answer is that climate has a visible route out. He does not claim the technology exists — "I'm not saying the technology currently exists that allows us to build, you know, some kind of power station that's clean" — but he believes it can be discovered, and says human history makes that undeniable. AI, he argues, is "completely unprecedented and a category different."
"And I am not pessimistic. I am just accepting of the truth of this issue, which is nobody can stop it. Nobody can control it."
Which brings him back to the only question he thinks remains, and he concedes it is not really an AI question at all. Everyone dies, and soon. The days between now and then are the thing you decide about: how to spend them "in a way that when you look back, you feel like you did the right thing." He does not claim to manage it — "I don't remotely do that every day. You know, I lose my temper. I make mistakes. I say things I shouldn't." Whether the end comes from a robotic chokehold three years from now, or in 40 years, or in 200 because AI has dissolved all your health problems, he says the question and the answer do not change.
Bartlett leaves the disagreement standing. Kisin closes on the one thing he says anybody controls: "what do you do when you leave the studio?"