Michael Lewis and Tom Lee took center stage at a podcast recording in New York City on Tuesday, speaking with SoFi’s head of investment research Liz Thomas for her program The Important Part. During a broad discussion that touched on Lee’s views about flash-frozen technology and Lewis’s dinner with OpenAI CEO Sam Altman regarding Sam Bankman-Fried, the two prominent finance figures examined whether the ongoing software stock selloff was escalating into a more severe situation. Their demeanor was serious yet witty.
The pair analyzed a market marked by severe turbulence, where software equities have plummeted significantly and artificial intelligence (AI) poses a threat to eliminate whole sectors. However, the most striking moment occurred when Lewis, the writer of The Big Short, presented a grim statistic about who truly profits in such conditions.
“Did you know Fidelity issued a report on its top-performing retail accounts?” Lewis inquired of the audience. “And they were all deceased clients.” (Lewis was alluding to a that discovered, in reality, the highest-yielding portfolios were those left untouched, whether from death or forgetfulness.)
Shortly afterward, Lee referenced a study showing that of 40,000 stocks that have launched or been spun out since 1974, 90% dropped over 50%, with most eventually becoming worthless. “So essentially, the bulk of stocks basically went to zero.”
FOMO or death?
The story highlighted a key theme of the night: in a market driven into a frenzy by fear of missing out (FOMO) and algorithmic trading, inaction is frequently the better approach.
“The takeaway isn’t to die,” Lewis clarified wryly. “Don’t trade excessively.”
Tom Lee, Fundstrat’s head of research, reinforced this perspective with his firm’s data. He observed that while institutional investors have compressed their investment timelines to just days—or in some instances, holding shares for “about 40 seconds”—retail investors are actually “making the right moves” because they operate with “permanent capital.” Being physically deceased, they cannot pull their money from the market. Unlike hedge funds that rotate positions based on daily profit and loss, individual investors hold their assets steady.
“As you know,” Lee addressed the audience, seemingly referencing high-frequency trading dynamics, as discussed in Lewis’s book , “the typical stock is held for roughly 40 seconds. So most of these major hedge funds … one or five seconds counts as a lengthy holding period. Many funds are essentially just cycling through stocks.” (Asset management executive has disputed such estimates, contending they relate solely to high-frequency traders and don’t reflect broader market activity.)
Lee acknowledged that Thomas had raised a valid point in her question, however.
“There’s something unusual happening this year. Suddenly, numerous stocks and industries are beginning to contract,” Lee stated. “Take the software sector, for example—it’s experiencing declining demand and a revaluation of its offerings, and multiple research reports now indicate that agentic AI or AI solutions are beginning to supplant conventional software.”
The contraction is substantial too. an iShares ETF monitoring software equities has hemorrhaged approximately $1 trillion in the past seven trading sessions.
Lee indicated he interpreted this as evidence of AI’s efficiency and a positive long-term development, arguing that software spending is decreasing because AI is handling those functions instead, and tech employment is lower now than in 2022, when ChatGPT debuted—and college graduate unemployment lagging non for the first time in living memory, if ever.
Lewis, however, said he detected parallels to the dot-com bubble. He cautioned that investors are once again “mistaking the technology for corporate profits,” presuming that AI’s transformative nature will automatically generate stock market gains.
“It could actually be a mechanism for diminishing corporate profits,” Lewis contended, implying that many current market leaders might ultimately “crash down.”
The discussion shifted to fundamental threats facing other asset classes too, illustrating a scenario where even “safe” investments could potentially become worthless. Lee speculated that Bitcoin might be made obsolete by quantum computing or even AI itself—if AI chooses to operate its own “validation language” and circumvent human cryptocurrency systems entirely. Even gold, which Lee calculates as a $35 trillion asset class, faces devaluation risk.
“There’s a million times more gold beneath the surface than above ground currently,” Lee said, arguing that if gold becomes too costly, the Magnificent 7 would simply enter the gold mining sector, “since you might as well excavate for gold.”
Amid the bubble talk and potential asset implosions, Lewis disclosed he has adopted a defensive stance—namely, the “Armageddon trade.”
“When I hold [gold], I believe I’m betting on fear,” Lewis confessed, acknowledging his position in the commodity, which he didn’t recommend listeners replicate. “I see no reason not to be fearful. And I think being long on fear is sensible at present.”
For the investors present who were still alive, the conclusion was paradoxical: The market is hazardous, technology is devouring profits, and the optimal survival strategy may be to imitate the deceased—the only dependable capital is permanent capital.