Sam Bankman-Fried Tells Teen Podcaster There’s a 25% Chance

Sam Bankman-Fried Tells Teen Podcaster There's a 25% Chance

The call began with a kid’s voice and the man who once ran one of crypto’s biggest exchanges. For a few minutes a prison line turned into a public thinking session about extinction. I heard Sam Bankman-Fried tell a 14-year-old podcaster he assigns a 25% chance that artificial intelligence will wipe us out.

I’ll be blunt with you: that sentence is a loaded grenade. You know the basics—FTX collapsed after billions of customer funds were used for risky bets, SBF got a 25-year federal sentence and faces $11 billion (€10.2 billion) in forfeiture—so you don’t need me to rehearse the headlines. What matters is what a disgraced, once-influential tech investor saying “25%” does to the conversation around AI, power, and who gets to define safety.

He told a teenager he’d put the odds at 25% — and then explained why.

What struck me first was how casually SBF framed probability. He didn’t reach for theater; he tried to sound like someone making a reasoned bet. He told Eli Goldfine that his estimate was “less than 50%, more than 10%,” landing around 25%.

What did Sam Bankman-Fried say about AI risk?

SBF argued that current models already reach a form of general intelligence in practical terms: you ask a model a question and you get an answer comparable to a smart human. He predicted “piece by piece” adoption over five years and warned the survival odds hinge on a peaceful relationship between humans and AI.

He also claimed that AI could “get bored of humanity and wipe it out.” Say what you will about the messenger, but when someone who lived through rapid, catastrophic collapse talks about societal fragility, you should at least pay attention.

You saw the investor math in his explanation of value and control.

He reasoned from the size of sectors: if sectors are worth trillions, frontier AI labs becoming trillion-dollar firms is a conservative bet. That’s not a moral argument; it’s an economic one.

SBF said he put $500 million (€465 million) into Anthropic and implied the scale of the prize justifies concentrated bets. When someone who helped turn FTX into a fountain of cash for risky strategies describes opportunity, you hear echoes of what happened with crypto. The defense sounded like a house of cards when you remember how that played out.

Is there a real chance AI could destroy humanity?

Experts disagree on probabilities, but SBF’s number matters because he’s staking reputation—and his prison years—on it. For people trying to prioritize safety, a 25% public estimate from a high-profile investor forces policy makers and funders to recalibrate urgency, even if you think the percentage is too high.

He talked to a 14-year-old on X and told him FTX’s philanthropy would have justified the rest.

He said the FTX Future Fund justified what he did—despite the $10 billion (€9.3 billion) of fraud that crippled the very causes it supported. That claim collapsed under its own weight: charities tied to FTX lost grants when the underlying accounts were tainted.

If you follow AI funding, you’ll recognize the pattern: money plus absolutist utility claims attract people and institutions. Anthropic, Claude, the myriad labs and policy shops—some tied to effective altruism—are all part of that ecosystem. Investors and idealists can be drawn in the same way: a moth to a flame.

Why did SBF speak to a 14-year-old?

Because platforms like X and podcasting reward unusual combinations—badged founders talking to improbable hosts get attention. Eli Goldfine is precocious and produces serious questions; that pair of novelty and sincerity is the kind of optics SBF, even behind bars, can use to shape a narrative.

I’ll say this to you plainly: I don’t trust reputation alone as a measure of expertise. SBF knows bubbles and how to sell inevitability. He also knows how to position moral ambition as a justification for risky behavior—he told Goldfine that without the fraud FTX “would have been worth it.” That’s a hard sell to the NGOs who were left sorting through the wreckage.

What should you take from this? Watch where influence flows. Track which labs and funders echo the same value frames he used. Read the DOJ filings and Reuters coverage, follow Anthropic and Claude, and watch how investment narratives shape safety conversations.

Subscribe to our daily AI newsletter
The 5-minute AI newsletter. The stories that matter, in your inbox every weekday morning.
Subscribe Now

If you want my practical read: take his estimate seriously as a rhetorical move by someone who knows how to make markets and narratives amplify each other. Watch policy, money, and lab governance where they intersect. And ask the people setting those rules what they mean by “the greater good,” and how much they’re willing to risk to get it—because if SBF’s odds are right, the rest of us don’t get a vote, do we?