Bring Back Web 1.0

Who gains from search’s demise—and can the cozy old internet become a form of resistance?

The internet is gradually ceasing to be a place for people and is turning into infrastructure for digital agents. Media are losing audiences, websites are losing their purpose, and knowledge is becoming a depersonalized synthesis produced by someone else’s algorithms. Why the “ten blue links” may be the last symbol of a human web, who benefits from the death of search, and whether the old, cozy internet can unexpectedly become a form of resistance — a ForkLog analysis.

Internet evolution or user degradation?

It’s hard to miss how fast the digital environment is changing with the rise of artificial intelligence and the advent of Web 3.0. Not long ago, internet users were mostly authors and commenters. Now some still sit in social networks and periodically ask: “Hey, does anyone see my posts? Leave any reaction if you do,” or “Folks! Where did everyone go, why is my feed only bot-generated content?” The days are gone when you could type a few words into a search bar and spend days opening links and reading something more or less meaningful on topics of interest. 

The new internet kills our sense of adventure; we stop feeling like pioneers, researchers, detectives, seekers of truth. We no longer spend hours at work sifting through and consuming gigabytes of information in the hope of stumbling upon something useful. People online talk to language models. And they get uniform, standardized, sparse answers in a ready-made — but usually not the best — form. 

Ideologues and marketers pitched Web 3.0, metaverses and AI as technologies of human liberation. We’re now at a stage where the user has already been turned into a client. But the day doesn’t seem far off when we’ll see a truly post-human internet. Against this backdrop, a nearly nostalgic question arises: might the new hero be the one who manages to build another internet — roughly as it was in the early 2000s? 

A caveat: in this article we touch on Web 3.0 and, to a lesser extent, Web3. While these concepts take different approaches, both aim to create a more capable internet, offering their own ways to solve current problems. Web3 emphasizes returning control over data and digital identity to users via blockchain technologies, while Web 3.0 focuses on making the internet more intelligent and efficient by reusing and interlinking machine-readable data across the network. 

R.I.P., Ten Blue Links

At Google I/O 2026, the company effectively made clear that results pages will no longer be just a file of links. The tech giant’s official blog says AI Mode has already become its most powerful search mode and surpassed 1 billion monthly users. If Google Search once answered by asking “which pages match the query,” it now acts on the interpretation of “what exactly the person wanted to know and how best to explain it.”

“We are bringing the cutting-edge capabilities of our model to Search with new AI features that let you use agents just by asking a question. We’re also introducing a new AI-powered search box — the most significant update in more than 25 years,” said Google Search Vice President Elizabeth Reid.

The global tech corporation has already decided for you and unveiled the concept of search AI agents. The company’s release assures that “you’ll be able to easily create, customize, and manage multiple AI agents to handle a wide range of tasks right in Search.” Nothing wrong with that — as long as you’re not under the power of invisible minions and they work for you. But the next paragraph of the release says: 

“With information agents, you’ll always be up to date on what matters most to you. Your agent will intelligently analyze everything on the web, including blogs, news sites, and social media posts, as well as our freshest data, like real-time information on finance, shopping, and sports, to monitor changes related to your specific question.”

In other words, you’ve been cut off from analysis. Which raises the question: “What next, are you going to eat for me too?”

We now have a “search engine with executive brains,” in which AI agents don’t just find information; they formulate clarifying queries, collect results, rank them, and deliver a finished answer or action. The user asks in natural language, almost as to a person: long phrases with context, clarifications, and follow-ups. Unlike the old “keywords — list of links” mode, this search tries to hold a dialogue, remember previous turns, and respond not with fragments but a coherent explanation. Yes, a machine is chewing information for us.

This shift aligns with broader market dynamics: according to Ahrefs data, the presence of AI Overviews is associated with a 34.5 percent drop in average CTR. Later analysis by Search Engine Land and Seer Interactive shows that when AI-generated answers appear, organic clickable traffic can fall by tens of percent, and users generally click less even outside those blocks. 

Against this backdrop, Google’s search system is clearly turning from a navigation interface into a layer of interpretation and delegation. Media outlets were among the first to feel the impact. Their task in dealing with platforms is changing: it’s less about appearing in results and more about being a source the system draws on to form an answer. For publishers, AI Mode primarily brings the risk of traffic loss, brand dilution, and dependence on someone else’s reading of their content. When users get a finished answer inside Google’s interface, they click original materials less often, so newsrooms lose visits, ad impressions, and the chance to keep readers on their own sites.

Google now fully decides which sources to show, how to summarize them, and in what format to present the answer, while media effectively become suppliers of raw material for someone else’s product. For journalism, that means a loss of control and influence. 

In parallel, an infrastructure for digital commercial agents (agentic commerce) is taking shape. The open Agentic Commerce Protocol already describes how they can make purchases, transfer payment tokens, and act on a buyer’s behalf. 

The problem of a radical change in search runs deeper than SEO metrics and sites becoming useless for failing to appear in results. When synthesis is done by a machine, the question of which sources it relies on moves from technical to political — especially given that decentralization of the internet, for various reasons, didn’t happen.

Google, buy me a hat

The balance of power among web users has truly shifted over 30 years. Those who used the web mainly as a super-dump of diverse information are now out of luck; their chances of finding gold dust and diamonds in tons of hyperlinks are close to zero. Centralized Web 2.0 not only seized your data but also became a big mommy who gently and lovingly says: “Eat what you’re given!”

Content creators so clogged the internet with their output, priceless opinions, advice, and simulated conversation that they stopped reading even themselves. LiveJournal died — and so be it; Twitter, that is, X, will die too. Everyone’s already overfed on social media, regularly going on digital detoxes and rehabs, starting to read paper books again! They are the offspring of that very Web 2.0 “mommy,” who feel duped yet still cling to some agency, launch Telegram channels, but can’t say anything new or interesting there — because it stopped being Googleable. 

Don’t take this for crotchety grumbling. Users didn’t ruin the internet; bloggers aren’t to blame at all. The internet became too big for a human mode of navigation. The real need for a systemic change has formed. 

And what do we become as prospective clients of Web 3.0? Buyers. But not like at a market, where you also look and sniff before you commit. Today’s internet trains the ideal customer who, instead of searching and transacting, sets a goal, and the AI agent takes over the search, comparison, choice, and payment.

The user gives a text or voice command, for example: “Buy the cheapest airline tickets to Rome for the weekend, a hotel no more than 100 euros per night, Wi-Fi required.” The AI agent independently scans marketplaces, booking sites, and aggregators. It either proposes a ready option for approval or immediately places the order using the user’s linked payment details.

Robots are working on the selling side, too. What is the person doing in the meantime? Have they freed up time for art, science, philosophy? In utopia — yes. In reality, without the opportunity and need to analyze, search, compare, and verify, we’ll quickly lose those skills. On top of that, the explanatory interface of new search engines inevitably expresses someone’s selection logic and thus imposes a particular picture of the world. 

If Web 1.0 gave access to information and Web 2.0 made everyone produce it, then Web 3.0 will spare humans from interacting with it at all. What are writers, journalists, editors, researchers, and readers to do in a system where the very principle of search has been “broken”?

It seems we need a counternet — a different space where sources remain more important than a synthesized answer, and where verification, accuracy, accessibility, and diversity of information stay more valuable than speed. Here the “old web” could become not a nostalgic indulgence for elders, not a rollback to the past, but a model of resistance and a new competitive arena.

Simple, link-driven, surveyable

The early web was more fragmented, and the total packaging of information into a single answer was absent. For some, that was less convenient. But the ability to research on your own, to see sources, follow them, compare versions, extract knowledge, and produce something new — that’s what there is to love about the cozy, lamp-warm Web 1.0.

Is that sufficient reason to consider — culturally, ideologically, and economically — creating an alternative “new old internet”? Quite. And many have not only thought about it, but started moving away from centralization, platforms, advertising, and bots.

The more actively digital agents act instead of us, the more valuable a web designed for human attention becomes. In academic, legal, scientific, and analytical circles, demand for verifiable, independent sources will grow as mass search keeps moving toward AI answers. That’s another reason to develop what we have conditionally called the counternet. This is best done without romanticizing the old internet. A literal return to Web 1.0 is impossible — and hardly desirable in full. 

A counterculture of life outside platforms and AI search already exists. Today it’s represented by several movements: IndieWeb, Small Web, Cozy Web, and lesser-known but kindred efforts. These initiatives aren’t a “new internet” in an infrastructural sense, but they try to return the web to a human scale: personal domains, small sites, direct links, manual navigation, and author control over their content. Their existence confirms demand for alternative web models and looks like an economic argument for building them.

At the same time, a return to Web 1.0 is unlikely to be a mass scenario. Most users will always choose convenience, speed, and delegation. AI agents do save time and remove drudgery. But precisely for that reason, the human internet may become a new form of “luxury” — a space without algorithmic noise, endless recommendations, and automated content. Not the main, big internet, but something like a digital preserve.