As the New York Times wrote in its review of Going Infinite, Lewis’s book about disgraced crypto founder Sam Bankman-Fried, the writer always “makes sure to give you an unsung hero to root for.” Hollywood has often found this a winning recipe for film adaptation. The renegade baseball general manager in Moneyball, for instance, was later played by Brad Pitt, while the characters in The Big Short were portrayed by Christian Bale, Ryan Gosling, Steve Carell, and Pitt again, among others. “Even Lewis’s first book, Liar’s Poker, which recounts all kinds of bad behavior on Wall Street, is structured around a young man named Michael Lewis,” the Times notes, with that underdog hero leaving the “absurd money game” of finance to become a journalist and celebrated author.

Altman, the influential and controversial Silicon Valley billionaire, transformed the economy by releasing ChatGPT and unleashing an AI revolution in November 2022. Almost a year later, he was fired by his board of directors for losing their trust, before rejoining over the course of one dramatic weekend. Along the way, he was sued by his former cofounder Elon Musk and became a lightning rod executive for the hype and fear surrounding the AI story. He could be a perfect Lewis character, and Lewis recounted to The Important Part how he met with Altman to discuss Bankman-Fried.

The challenge

The two shared a dinner while Lewis was researching Going Infinite, because Lewis had sought out Altman to discuss Bankman-Fried, given that the FTX founder had purchased a 20% stake in OpenAI rival Anthropic using customer funds. “But it turned out he had a little bit of an agenda,” Lewis told SoFi’s Liz Thomas, head of investment strategy and host of The Important Part. “And his agenda was, he had all these people who wanted to write his biography, and he didn’t really want his biography written, but he thought if he could just pick one of them, it would get everybody else off his back. So he wanted to talk to me about some of these people.”

Lewis, known for his skepticism regarding financial and technological hype, pushed back with a question that cuts to the core of the current artificial intelligence frenzy. “I said, ‘Well, if your machine is so smart, why don’t you have your machine write the biography?’” Lewis recalled suggesting that Altman dump all his Slack channels, emails, and personal data into the model to let it tell his story.

Altman’s response was surprisingly candid: “He said, ‘It’d be a really bad book.’” When pressed on when the AI would be capable of such a feat, Altman estimated, “Maybe a couple of years.” This led to a unique challenge between the author and the tech mogul.

When Lewis and Altman agree, at some point in the future, that ChatGPT is finally smart enough to write a good book, Lewis said he wanted to “challenge it to a contest. I will come back in, and we will pour everything into it. We’ll let it write its book, but I’ll do my book alongside it. And we will just compare whose book is better.”

Several books about Altman and OpenAI are already in print, including Keach Hagey’s The Optimist and Karen Hao’s Empire of AI. Columbia law professor Tim Wu reviewed both of these for the New York Times, writing: “Hey ChatGPT, Which One of These Is the Real Sam Altman?” The works by the two journalists offered complementary portraits of OpenAI’s “notorious” cofounder, he added, saying that both books suggest Altman has created a real-life version of the ethics thought experiment, or the “paper clip problem.” The idea is that an AI, tasked only with making as many paper clips as possible, would transform everything around it into a giant paper clip manufacturing facility. They seem like good books, in other words.

Could 2026 be the turning point for AI?

There was a broader tension between Lewis and co-panelist Tom Lee, head of research at Fundstrat, discussed during the event. According to Lee, the very metrics that look like “bad news” are actually confirming a massive leap in economic efficiency. “Technology employment has shrunk in the last three years since the introduction of [ChatGPT],” Lee observed, noting that for the first time, college graduates’ unemployment rate is now higher than that of non-college-students of the same age.

While this contraction in the workforce is painful for workers, Lee argued that it’s the textbook definition of technological success. “That’s productivity for the economy. This is how you measure productivity, is that you produce more output with less people,” Lee said. “So that’s proof that AI is productive.” And so is the current swoon in markets, he added, noting the period of volatility where software stocks have been “down a ton.”

Lee attributed this repricing to a fundamental shift in how corporations spend money. Companies are “spending less on software,” creating a “margin story” where “agentic AI products or AI products are starting to replace traditional software”—even more proof of how productive AI is. Lee predicted that 2026 will mark the era where the market sees “the other side of the payoff from AI,” a shift that is “creating winners and losers” rather than lifting all boats indiscriminately.

Lewis said he agreed that the AI boom reminds him of the dotcom bubble in some respects because of the winners-and-losers dynamic: “There’ll be some things you were really happy you bought that are still going to go up. There are going to be things that come crashing down.” The other big similarity, he argued, is that people are “conflating the technology with corporate profits. Like, just because the technology is transformative doesn’t mean that people are going to make a lot of money from it, or that everybody’s going to make a lot of money from it.”

Will AI take your job?

For now, Lewis said he isn’t sweating the competition. He admitted to testing AI during his writing of Going Infinite, asking it to write a book about Sam Bankman-Fried to see what it could produce. The results were unimpressive. “All it knows is what’s on the web … It hasn’t gone out and interviewed people. It hasn’t seen and smelled and tasted and all that,” Lewis explained. He said he didn’t see AI taking his particular job anytime soon. “I don’t really feel yet, honestly, a threat from AI.”

Lewis noted the hype from AI is extraordinary—and extraordinarily disturbing. Many AI CEOs are alternating, he noted, between two doomsday predictions. “The people at Google, the people at OpenAI will very cheerily … say out of one side of their mouth that, you know, [AI] might kill us all, so we ought to be really, really careful; we need to think hard about it. But nobody seems to be doing anything about that … At the other side of their mouth, they’ll say, ‘In 18 months it’ll be able to do everything a human does, but better.’”

The second prediction is really worth taking seriously, Lewis argued: “If in 18 months it can do everything that a human can do, but better, what on earth is going to happen to this country?” Lewis noted the anger over affordability and the economy that helped elect President Donald Trump is a real problem that is far from resolved: “We’re literally living with the politics of anger for a stretch. The anger is going to go through the roof” if AI takes that many jobs in just a year and a half.

Lee offered a philosophical perspective, repeating some of his recent research about disruption and the odd tale of “flash-frozen” foods, as Fortune previously reported in January. On the Prof G Markets podcast that month, Lee noted that the third great labor-shortage era in U.S. economic history began in 2018 and would last to 2035, and offered up the invention of “flash-frozen” foods in the 1920s as a parallel to AI.

Fundstrat research indicates that farming went from 30% to 40% of the workforce before flash-frozen food to 2% to 5% afterward, but enough jobs were ultimately created to account for this reduction in employment. “So, like, 95% of farmers were replaced,” Lee told Lewis and Thomas. “That created time and innovation … The economy prospered, even though an economist would have told you in 1930 that we’d have another depression.”

There was a less successful transition in the 1990s, he added, “with China taking over manufacturing.” Lee noted that it “gutted” entire states. “People call it distributional consequences, but that was poorly planned because they didn’t think about how to find new work for these people. So I think it is, like Michael said, pretty important to be thoughtful of the distributional consequences of what’s happening.”

Lewis was bowled over by Lee’s research, which included the revelation that an early version of Goldman Sachs was a venture investor in Clarence Birdseye’s flash-freezing method; Birdseye studied as an ornithologist (as well as a taxidermist, fur trader, hunter, and fish lobbyist). Lee explained that Birdseye was studying birds when he was served fish by Alaska’s Inuit tribe, and was shocked at how fresh the fish tasted, because it was flash-frozen on ice. “His name is Clarence Birdseye, and he studied birds,” Lewis remarked, shaking his head.

Lee noted that he’d been doing deep research on such topics since he began his career on the research team at Salomon Brothers, the same (now extinct) Wall Street firm where Lewis began his career, as documented in Liar’s Poker. “I had no idea you were at Salomon Brothers,” Lewis said. “Salomon Brothers had the smartest people in the research department. They were just the smartest people I’d ever met.”

“Yeah, I read your book,” Lee responded. “But I wasn’t a master of the universe.”