In October 1945, a young British writer called Arthur C. Clarke published an article in a magazine called Wireless World with a strange idea. Put a satellite about 36,000 kilometres above the equator, he said, and it will circle the Earth once every 24 hours, so it will seem to hover over one spot. Place three of them around the planet and you could send radio signals anywhere. Almost nobody took it seriously. In 1965, twenty years later, the first commercial satellite of exactly that kind, Intelsat I, went into orbit. Clarke had called it.

For every Clarke, though, there is a Clifford Stoll. In February 1995, Newsweek published his essay "The Internet? Bah!", which dismissed online shopping, working from home and internet journalism as "baloney". He later admitted it was a howler. Nearly a century earlier, in 1900, Ladies' Home Journal asked a writer to predict the next hundred years. John Elfreth Watkins Jr. correctly foresaw ready-cooked meals delivered to your door, but he also promised that the letters C, X and Q would vanish from the alphabet. They are all still here.

So are experts any better? The psychologist Philip Tetlock spent nearly twenty years collecting 28,000 predictions about politics and the economy from 284 experts. Their forecasts were often only slightly better than chance, and, as one much-repeated line puts it, about as good as a dart-throwing chimpanzee. The most confident experts, the ones with one big theory about how the world works, did worst. The best forecasters were people who drew on many sources, doubted themselves, and changed their minds.

Which is probably the most sensible advice about the future anyone can give. Nobody has a crystal ball. The skill is not predicting tomorrow perfectly but holding your predictions loosely, checking how often you are wrong, and being ready to update. So the next time someone tells you confidently what the world will look like in 2050, it may be worth asking one polite question: what did they get wrong last time?