Sam Altman's First Company Failed. Nobody Remembers That Part.
In 2005, a nineteen-year-old dropped out of Stanford and started a company called Loopt.
The idea was location sharing — you could see where your friends were on a map, in real time, on your phone. It was early. The iPhone didn't exist yet. The App Store didn't exist yet. The infrastructure for what Loopt was trying to do was, in most meaningful ways, not there.
Loopt was accepted into the first batch of Y Combinator. Paul Graham gave them $6,000 and a few months of office hours. They built, they shipped, they grew — and then they didn't grow enough. Location sharing turned out to be a feature, not a company. Facebook would eventually absorb the idea into its own product and render Loopt irrelevant.
In 2012, Loopt was acquired by Green Dot for $43 million. Not a failure by most definitions. Not a success by the definitions that matter in Silicon Valley. The founders made some money. The company disappeared.
The founder, Sam Altman, went on to become president of Y Combinator.
This is the part of the story that gets skipped. Paul Graham — who had seen thousands of founders by the time he chose Altman — did not choose him because Loopt had worked. He chose him because of how Altman had thought about Loopt while it wasn't working. In a 2009 essay, Graham named Altman one of the five most interesting founders he had ever met, alongside Steve Jobs and the co-founders of Google. Altman was twenty-four. His company was still alive but already losing.
"On questions of strategy or ambition," Graham wrote, "I ask: what would Sama do?"
Altman ran YC from 2014 to 2019. Batch sizes went from 40 companies to 200. The programme became, definitively, the most important place a startup could begin. He did not do this by being the smartest person in the room. He did it by being the person most willing to think about what YC should become, rather than what it had been.
In 2015, he co-founded OpenAI as a nonprofit. The mission was to ensure that artificial general intelligence benefits all of humanity. Elon Musk was there. Peter Thiel was there. The money was serious and the idealism was genuine, which is a combination that rarely survives contact with reality.
It didn't, entirely.
OpenAI became a "capped profit" company. Microsoft invested $13 billion. The nonprofit wrapper remained but the commercial engine underneath it grew to a scale the original structure had never anticipated. Every step of this transformation was contested. Safety researchers left. Musk sued. The mission statement stayed on the website while the business underneath it evolved into something more complicated than any mission statement can contain.
And then, on a Friday in November 2023, the board fired him.
The stated reason was that he had not been "consistently candid" with the board. No specifics were given. The board had four members. One of them, it later emerged, had been feeding information to a journalist for months. Another had written academic papers arguing that AI development should be slowed or stopped. The decision was made over a weekend, announced without warning, and immediately produced the most visible employee revolt in the history of the technology industry.
Seven hundred of his seven hundred and seventy employees signed a letter saying they would quit if he was not reinstated. Microsoft, which had just committed billions to the company, made its position clear. By Monday, the board that had fired him had dissolved itself. By Tuesday, Altman was back.
The media covered this as a drama. It was a drama. But underneath the drama was something more interesting — a test of what happens when the people who built something decide that the person nominally in charge of it does not, actually, control it.
Altman passed the test. Not because he fought back. Because seven hundred people chose him over the institution.
He is not, in the conventional sense, a sympathetic figure. He runs the most powerful AI company in the world. He has more access to capital and compute than any individual in history. He gives interviews in which he speaks about existential risk with the calm of someone who has already priced it in. The media finds him slippery. Critics find him convenient. The people who work for him, apparently, find him worth following into uncertainty.
FreeMalta is an Official OpenAI Select Partner. We built this partnership not because we agree with every decision OpenAI has made — nobody reasonable agrees with every decision any organisation has made — but because we believe that the technology being built there is real, that the people building it are serious, and that Malta's businesses deserve access to what it can do.
Sam Altman's first company failed. He was recruited to run YC because of how he thought about failure. He built OpenAI because he believed something important was happening and someone had to try to shape it responsibly. He was fired and came back in four days because the people around him decided the alternative was worse.
You do not have to like him. You do not have to trust him. But you should probably stop dismissing him.
The nineteen-year-old who dropped out of Stanford to build a location sharing app that didn't work is now the person making decisions about the most consequential technology of our lifetimes.
That is either terrifying or reassuring, depending on what you think of him.
Both reactions are reasonable. That, too, is rare.
