Type a question into Google and imagine the familiar obstacle course: an advertisement, a page that repeats your question, then an introduction that takes an astonishing amount of time to arrive nowhere. This is a caricature, not a measurement of every search result. Its usefulness lies in how readily we recognise the situation. We wanted to understand something. The page wanted us to arrive.
Those two objectives can overlap. They are not the same objective.
Calling search prehistoric is a provocation. The genuinely outdated habit is not typing into a box. It is treating the order of the results as an order of knowledge, as though a position on a screen could settle whether a claim deserves our trust. The interface offers a route through information. We quietly promote it to a judge.
Visibility is not authority
A result can be relevant to the words in a query without answering the problem behind them. A page can be easy to find without being written by someone competent to address the issue. A statement can be confidently presented without explaining the conditions under which it holds.
These distinctions sound obvious when written down. They become less comfortable when we are busy, tired, or simply looking for a quick answer. The first plausible result supplies a convenient stopping point. Convenience then starts doing the work that evidence should have done.
The criticism is not that every prominent result is wrong. That would merely reverse the same mistake. Ranking is not proof of truth, but neither is it proof of falsehood. It is a feature of discovery. The intellectual work begins when we examine what has been discovered.
The reader and the visit
Search engine optimisation is not inherently dishonest. Making a useful document discoverable can benefit its author and its reader. Clear structure, relevant wording and accessible pages are not crimes against civilisation.
The tension appears when attracting the visit becomes more important than serving the person who visits. Consider a hypothetical article built around a practical question. Its headings promise an answer, its introduction repeats the problem, and its conclusion leaves the important exceptions untouched. It has successfully intercepted demand for an explanation without supplying much of one.
The reader measures success in understanding. The publisher may also care about traffic, subscriptions or sales. These interests can coexist, but the interface does not guarantee that they coincide. A polished page can conceal this divergence as easily as a clumsy one can reveal it.
That is why more browsing does not automatically mean more learning. Ten tabs can contain ten versions of the same unsupported assumption. We have multiplied the windows, not the evidence. A browser full of open tabs is not, unfortunately, a research department.
What search still does well
Google remains useful for reaching a known website, locating a specific document or finding material to investigate. Dismissing all of that as useless would make a satisfying slogan and a poor working method.
The important distinction is between finding and judging. If the task is to reach an organisation’s published policy, a search result may get us there. Whether that policy applies to our situation, whether it is current and whether our interpretation is sound are different questions. The journey to the document does not answer them.
The same distinction matters for technical guidance. A page may contain a workable answer for one version, one environment or one set of assumptions. Applying it elsewhere requires more than copying the most visible paragraph. Context is part of the answer, not optional packaging around it.
Search becomes misleading when we ask it to perform this judgment without noticing that we have done so. We stop asking what supports the claim because the claim has already passed the much easier test of appearing in front of us.
A chatbot is not an escape hatch
Replacing the results page with a conversational answer does not automatically solve the problem. It changes the presentation. Instead of several links competing for attention, we may receive a coherent paragraph that seems to have finished the reasoning on our behalf.
That fluency can be useful. It can also make uncertainty harder to see. A response without inspectable support does not become reliable merely because it is well written. It may deliver the same old problem with better typography and a more accommodating tone.
The sensible comparison is not between an allegedly obsolete search engine and an allegedly infallible assistant. Neither format deserves unconditional trust. The relevant question is whether the tool helps us reach evidence, understand its scope and recognise where the answer stops.
An assistant can help organise questions, explain terminology or identify what is missing. Those are valuable functions. They do not remove our responsibility to distinguish an explanation from a demonstration, or a plausible synthesis from a claim supported by the underlying material.
From results to reasons
A better information habit begins by clarifying the task. Are we looking for a particular document, a definition, a practical procedure, or a judgment under uncertainty? These are different jobs. Asking a vague question and accepting a neat answer can hide the fact that the job was never properly specified.
Next comes provenance. Who is making the claim? Are they describing direct knowledge, interpreting someone else’s work, selling a service, or expressing an opinion? None of these categories automatically settles the matter. They tell us what kind of statement we are reading and what support it needs.
Then come the limits. What date, jurisdiction, technical version or set of circumstances does the answer concern? What would have to change for it to stop being useful? An answer that names its boundaries may be less immediately comforting than a universal promise, but considerably more useful.
The effort should match the stakes. Finding a restaurant address does not call for the same scrutiny as acting on legal, security or financial guidance. The point is not to turn every search into a dissertation. It is to stop allowing convenience to decide, invisibly, how much scrutiny matters.
Faster access, the same old surrender
The strongest criticism of contemporary searching is therefore not that Google can never help us. It is that access has become so convenient that we can mistake arrival for understanding. We reach a page, recognise the vocabulary, and feel that the question has been dealt with.
Moving that reflex into a chat interface does not modernise it. The genuinely useful change is from collecting answers to examining reasons: who says so, on what basis, within which limits.
We have made it faster to obtain a response. Believing it because it is conveniently placed remains the prehistoric part.
Raffaele Di Marzio
All my “insane” books on cybersecurity and governance are here 👇 https://www.amazon.it/stores/author/B0FB47T6Q4/allbooks