About

Somebody famous predicts the future of AI roughly once a day. AGI in five years. Half of all jobs gone in ten. A million robotaxis by next summer. The predictions make headlines, move markets, and shape how everyone thinks about what’s coming. Then the deadline arrives, quietly, and almost nobody goes back to check.

That bugged us. So we built the scoreboard.

The AI Prediction Index collects predictions that are specific enough to be wrong. Each one gets a source, a date it was said, and a deadline. When the deadline passes, we look at what actually happened and hand out a verdict. No vibes, no grading on a curve. The rules are spelled out on the methodology page, and you’re free to disagree with any call, because every entry links back to the original claim.

One thing worth saying plainly: this site is not here to embarrass anyone. Putting a date on a claim takes nerve, and the people who do it deserve more respect than the ones who only ever speak in fog. Herbert Simon predicted in 1965 that machines would do any human job within twenty years. He was wrong, and he was still one of the great minds of his century. Landing on this list means you said something that could be checked. That already puts you ahead of most people talking about AI.