We’ve been using computers wrongly all this time, says 1440’s Guy Corbet.
For thirty years, we’ve accepted something rather odd. Whenever we wanted a computer to help us, we first had to learn how to help the computer. We translated ideas into keywords, navigated folders and menus, and stitched together answers from lists of links.
Sixty years ago, in the 60s, Star Trek imagined something quite different. Captain Kirk never asked the Enterprise computer to return ten documents about getting to warp speed. He simply said, “Computer, find a way…”. The machine’s job was to solve the problem. That always felt like the future. It felt like what computers were for.
What will people sixty years from now think about the way we used computers in the early twenty-first century? I suspect they’ll find it rather strange.
For us, the internet arrived and something curious happened. Computers did lots of the work, but so did we. We learnt keywords, search techniques, menus, folders and links. Google was extraordinary, but perhaps we’ve mistaken it for the destination rather than a stage on the journey.
Indeed, perhaps we’ve been doing it wrong for all these years. Not because Google wasn’t brilliant. It is. But because we’ve mistaken a work around for the destination.
Now AI is turning our assumptions upside down. Rather than asking people to adapt to computers, computers are finally beginning to adapt to us.
The remarkable thing about ChatGPT and similar tools isn’t that they break the rules. It’s that they finally follow them.
We ask a question. The computer gets on with solving the problem. It behaves much more like the machines we imagined decades ago.
For thirty years we’ve assumed the internet was teaching computers about the world. In many ways it was teaching us how to behave like computers instead.
We learnt to think in keywords rather than ideas, to optimise queries rather than express problems. We stopped asking computers what we wanted to know and started asking them what we thought they wanted to hear.
We guessed which words the computer wanted rather than simply asking the question we really wanted answered. We adapted ourselves to the machine. We assembled the answers ourselves. Very often, that was the hard work.
AI reverses that relationship.
Conversation replaces keywords. Searching becomes asking . Websites become evidence rather than destinations . Content is no longer created just for people to read. Increasingly, it is something AI reads, interprets and weaves into its understanding on our behalf.
That has consequences far beyond technology.
The computer is no longer simply finding information. It is assembling understanding. Organisations will increasingly be judged not simply by what they say, but by the wider body of information AI assembles about them. The question shifts from “can people find us?” to “what understanding will AI build when it does?”
Perhaps that’s why this transition feels so significant. The historic blip may not be AI and the shock of the new. The historic blip may be the thirty years in which we all accepted that using a computer meant learning to think like one.
In the future, people will look back on the Google era with genuine affection. It has transformed access to information. But they may also wonder why we spent so much time hunting through lists of links ourselves.
Perhaps Google wasn’t the culmination of computing after all. Perhaps it was the bridge between the old world and the one we’d imagined all along.
The internet taught a generation how to ask computers questions. Now AI is teaching computers how to understand ours. That’s how it should always have been.