The Death of Search
A while ago I stopped searching the internet the way we used to, and I think almost all of us are going to stop. Going back to the old way already seems absurd to me. I had to hold a question in my head that was worth the trouble, turn it into the words the engine understood, open half a dozen links and piece together a little from each until I felt I had found the answer, which usually opened three more questions. It was like rummaging through an old library, it always left room for interpretation, and it burned through the one thing you cannot get back, time, both mine and my company’s. Now I ask a machine and have the answer before I finish putting the question into words.
I am not a neutral observer. At Hellomatik we build agents that work across an entire company, I make my living betting on this happening, and I have bought a hundred Google shares. I have used these tools every day for years, and that is where I write from. I am not sorry that search is dying; what I want to describe is what it does to the mind of someone who stops searching, and where I fear it leads.
People read this as Google’s defeat. It is not. Its search engine is not falling, it keeps growing, only more slowly, and it answers you itself more and more before sending you off to the links. Google is in a different business, answering the question of the moment, fast and without complications. You do not ask it to do your taxes; for that you reach for a powerful model. You ask it when a shop closes or what a word means, which is most of what gets searched. For that question it does not need the best model, and it knows it. It came late to the race for the most capable model, so it is playing a different game, on far larger ground: what I called, in another essay, the AK-47 of intelligence, a cheap, sturdy tool that is enough for almost everything, against the few expensive, exquisite models, the "best or nothing" kind, as Mercedes puts it, kept for the hard problems.
This has happened before. Back in 2022, when you could search from a chat, Google’s stock took a hit and many of us stopped searching on Google. Not long ago it built artificial intelligence into its own search engine. I was among the first to notice, and then something I had not expected happened: those of us who had got used to asking a chat every stray question drifted back to Google for the quick ones, because it answers sooner and it is enough.
The shift shakes everything that lives off search, and the market is already pricing it in. On 28 October Google’s stock fell more than ten percent when it reported earnings, its first stumble in years, with search growing more slowly than investors wanted, and Reddit had dropped earlier when it warned that AI search was draining its traffic. It changes how a business gets found, and with it SEO and pay-per-click. The businesses that advertise on Google do not disappear, but they are forced to change shape. I would bet Google ends up putting advertising into its AI answers, on its lower tiers, and that a few strong players come out ahead while those who lived in the middle of the list are squeezed out. Whoever makes good content, on the other hand, can get the AI to name them when someone searches for something close to what they offer. Even the craft changes its name, from SEO to GEO, preparing your pages so that an artificial-intelligence model picks them out rather than leaving them to fight for a place in a list of links.
And the day the links stop drawing interest, Google will switch them off. There is a precedent: the "I’m Feeling Lucky" button, which took you straight to one result without the list. Young as I am, I still saw it every day not long ago, and Google has quietly let it fade; the links can go the same way. The website will be left as a last step, built around the call to action, because the AI will have done the sifting first, and if someone reaches it, it is because they really are interested.
Beyond the business, what unsettles me is what this leaves in the mind. The effort of searching shrinks to almost nothing; you ask yourself something and have the answer before you finish saying it. And that is where the risk I worry about most comes in, the carelessness of taking a wrong answer as true, carrying it into a decision or repeating it to a colleague without noticing the mistake.
And it brings me back to what I always end up at. Raw knowledge is worth less and less, because the barrier to reaching it keeps dropping. First it was being able to read, then living near a city with books, then owning a computer to search with, and now, for anyone with the means, it is enough to ask. Knowledge was always a matter of joining things from different fields, and a machine does that in a second. Knowing matters less now, and the judgment around it matters more, which is what you ask and what you do with the answer. My own defense, the one I use every day, is to distrust the first answer and put whatever matters to other models that judge it blind, without knowing which answer I lean toward; more than once they have preferred the one I would not have taken.
Who gets to speak has followed a similar path. At first a few spoke, the newspaper and the radio, then television, then anyone on a network, and now a machine that answers each of us privately. If we had asked someone in the year 2000 how we would get our news today, they would not have pictured an agent that answers or a feed of strangers weighing in on the day’s events; they would have pictured a television with a sharper picture, because the next revolution is very hard to imagine. I am afraid I am just as blind.
Even so, I will venture one part. I fear that communication between people will shrink and that content will become personal, made by a machine and tailored to each of us. Each of us will have our own version of the world, with people who do not exist advertising products in fake videos and the news told in the tone we like. The same assistant you hand your daily tasks to will also be the one telling you what is happening in the world, so there will no longer be a common place to check it from. The internet, built to bring knowledge and people together, will have ended up separating them, each into a world of their own, a black box we ask. I am well aware that this tailored world resembles the one I build; I tell myself mine is governed by a company over its own data, not by an app over a single person, but I know that is half an answer, because the very convenience I criticize works on what I build too.
About what comes after I only have a hunch, and I cannot prove it, but I believe it anyway: when finding costs nothing and being entertained costs nothing, they will stop being worth anything, and people will go back to wanting what they do not have, the analog, the human, the work of deciding for themselves, the way fashion revives what it once discarded. That is still far off; the first part, that search as we knew it will die, I take as certain. None of this is simply the way the world goes. It is a chain reaction, and the people building this future are as responsible for it as those who built the first search engines. The only thing I know to do about it, for now, is to keep the machine from taking the last decision off my hands as well.