Home

Sam Bankman-Fried Tells 14-Year-Old Podcaster There’s a 25% Chance His Days Are Numbered

“Sam Bankman-Fried talked to a 14-year-old podcaster and said there’s a 25% chance that artificial intelligence will destroy humanity” sounds like a Mad Libs sentence. Unfortunately, we live in a Mad Libs-ass world anymore, so it’s a real thing that happened—though your guess as to why is as good as anyone’s.

SBF, you’ll remember, is the disgraced founder of FTX, the massive cryptocurrency exchange that fraudulently used billions of customers’ money to make risky bets and was unable to fulfill withdrawal requests when the walls started closing in. He’s currently serving 25 years in federal prison and is on the hook for $11 billion in forfeited assets stemming from his fraud. So you could say he’s got some free time.

He recently spent it talking to Eli Goldfine, a 14-year-old podcaster and forecaster who—with no malice directed at him at all because he is a literal child and by all accounts seems very smart and precocious and certainly more productive than most teens—seems like they shrunk Lex Fridman in a dryer.

EXCLUSIVE: I'm 14, and I just interviewed Sam Bankman-Fried.

We talked about:

– Whether FTX was worth it– His p(doom)– Why he only put $500M into Anthropic– Whether he’d flip a 51/49 coin for his own freedom– How FTX almost became a political prediction market– Why he… pic.twitter.com/c34pQTd56t

— Eli Goldfine (@realTomBayes) November 20, 2026

Goldfine, to his credit, managed to get SBF to talk about a wide-ranging number of topics, from the downfall of FTX to his undying dedication to the not-technically-a-cult of effective altruism, to the permanently embarrassed former billionaire’s thoughts on the artificial intelligence boom. Bankman-Fried is certainly an interesting person to talk to about the ongoing AI boom, not only because he was an early investor in Anthropic, but also because he knows what it’s like to be at the forefront of a massive tech bubble.

Shocker, he doesn’t see AI that way, though. “You look at the biggest sectors, the sectors that drive parts of our economy, and they’re worth trillions to tens of trillions of dollars,” SBF said. “There’s every reason to think that over the next decade, AI is going to become one of those core sectors, if not even the dominant one of those.”

He reasoned that because there are so few frontier AI companies and economic sectors are worth so much, it’s basically a conservative estimate to assume the main labs will be worth trillions, which is not really an argument for utility and way more of a condemnation of how the economy works than he likely intended—though it is surprisingly and accidentally insightful if you apply that same reasoning to how he approached FTX and the crypto space—basically, there’s a lot of money here, someone’s gotta take it.

Asked about the potential for AI to pose an existential threat to humanity, SBF told the teen, “I would say, like, less than 50%, more than 10%. You know, somewhere in the 25% range.” He noted there are a lot of variables and differing opinions as to what accounts for a “catastrophic event,” but he did suggest that “over the next five years, we’ll see piece by piece, parts of our society be subsumed by AI.”

He also claimed that models are already at a form of “general” intelligence, in that you can ask them any question and “get an answer that is comparable to what you get from a smart, intelligent human.” Unclear how SBF knows that unless his prison has a Claude subscription, but that’s his understanding of the models at this stage.

SBF said our chances at survival will come down to whether humans and AI can develop a “peaceful coexistence” and avoid the risk that “AI will just get bored of humanity and wipe it out.”

Maybe SBF thinks he’d do better with the AI overlords because they would understand his “greater good” mindset that led to him illegally using customer funds and lying to investors and the public. He did tell Goldfine that if not for that little detail of the $10 billion worth of fraud, FTX would have been worth it because of the good done by the FTX Future Fund—seemingly forgetting that, because of that teensy weensy screw-up, he actually screwed over a bunch of the causes he was trying to help because they got stuck with fraudulent gains propping up their efforts.

Unfortunately for SBF, you typically don’t get credit for good intentions when you commit massive fraud to achieve it and end up hurting the people you wanted to help. But it seems the effective altruism math still gives him a different answer—which, if nothing else, should probably raise some red flags about the presence of EA actors all throughout the AI industry. It might be time to start asking those folks what the greater good is to them, exactly, and what they’re willing to sacrifice to achieve it.

Source: Gizmodo

Previous

Next