The stolen crystal ball

A viral story this week is celebrating Sam Bankman-Fried as “a genius at picking generational winners.”

The numbers are eye-catching. An eight per cent stake in Anthropic, bought for $500 million, is now notionally worth over $30 billion. A $200,000 cheque written to a small AI startup called Cursor, today worth $3 billion. Solana. SpaceX. Robinhood. The posts add it all up to $114 billion that the FTX bankruptcy estate supposedly left on the table by selling too early.

The conclusion making the rounds is that SBF was a visionary. Tragically undone by panicked lawyers.

This is, as politely as I can put it, nonsense.

But the wrong-headedness of this story is more interesting than it first appears. It rests on an assumption that almost every retail investor quietly shares: that the hard part of investing is picking what to buy.

It is not.

The hard part, the part that separates real wealth creation from cocktail-party stories, is holding on. And SBF, for reasons that have nothing to do with luck or skill, was structurally incapable of holding any of it.

Start with the obvious. The money was stolen. Every dollar that went into Anthropic, Cursor, and the rest was customer money he had no right to invest. The genius, such as it was, lay in the theft. Not in the stock-picking.

Now think about what holding actually requires. When Anthropic’s valuation wobbled, when Solana crashed from $260 to $8, when Cursor was a four-person outfit with no revenue, staying invested through all of that demands something specific.

It demands legitimate ownership. Capital that no one can force you to liquidate. Patience that comes from knowing the money is yours, and that no one will knock on your door and demand it back.

SBF had none of that. The FTX estate did not sell those assets at the bottom because the lawyers were stupid. They sold because they had to. When you steal billions from customers, the court will, quite rightly, seize your assets and convert them to cash to repay the people you robbed. The estate eventually recovered $18 billion and repaid creditors with interest.

That is the system working as designed.

Warren Buffett has said, in various ways over the decades, that the stock market transfers money from the impatient to the patient. But patience is not a personality trait you can summon at will. You can only be patient if your capital is legitimately yours, your time horizon is genuinely long, and no one has the legal authority to force you out.

There is also pure survivorship bias in this story. We see the Anthropics and the Cursors. We do not see the dozens of other Alameda bets that went to zero, because nobody writes viral threads about those. Spraying stolen money across a hundred venture bets and then cherry-picking the winners in hindsight is not genius. It is a lottery ticket bought with someone else’s money.

There is an Indian parallel worth drawing here, even if it sounds unglamorous. The investor who started a Rs 10,000 monthly SIP into a diversified equity fund in 2008 has lived through the global financial crisis, demonetisation, GST, the 2020 Covid crash, and every panic in between. Nobody writes breathless threads about her. She has no $30 billion entries to point to.

But her capital is hers. No court can seize it. No margin call will force her out. No bankruptcy estate will be liquidating her holdings on the worst possible day. That dull, structural certainty is what compounds over the decades.

You do not need to pick the next Cursor. You need to invest money that is honestly yours, in things you understand, and then do nothing for a very long time. Legitimate capital, patiently held, is the only edge that has worked reliably for a hundred years.

It will not go viral. But it is the only version of the story that ends with you still holding the asset.

Also read: How long is long-term?