The web still holds more useful information than ever, but getting to it now means clicking through cookie banners, dodging autoplay videos, scrolling past AI summaries and wading through pages written for search engines instead of people. None of this happened by accident. It is the result of a few business models stacking on top of each other over two decades, and once you understand them, you can route around most of the damage.
Ads and trackers became the business model
Most websites are free to read because advertising pays for them. That trade worked reasonably well when ads were simple banners sold against a page topic. The shift came when ads started being sold against people instead of pages. Programmatic advertising runs a split second auction every time a page loads, and the bidders want to know who you are, where you have been and what you might buy.
To supply that data, a typical news or recipe page loads scripts from dozens of third party companies: ad exchanges, analytics tools, session recorders, social widgets and data brokers. Each one slows the page and creates new ways to follow you across sites. Publishers often earn only a slice of each ad dollar, so they compensate by adding more ad slots, sticky video players and popups. The page gets worse, and the cycle repeats.
Cookie banners: a privacy fix that turned into friction
Consent rules in Europe, and a growing patchwork of state privacy laws in the US, require sites to ask before tracking you. In theory that gives you control. In practice, many sites use consent pop ups designed to make agreeing easy and refusing tedious, with a bright Accept button and a maze of toggles behind Manage options.
The tracking itself did not go away either. Google spent years planning to phase out third party cookies in Chrome, then reversed course. In October 2025 it confirmed that Chrome would keep its current approach of letting users choose about third party cookies and retired most of its Privacy Sandbox replacement technologies, including Topics and Protected Audience. For everyday users, that means the banners and the cookies both stay.
Search got gamed, then got replaced by answers
Search engines rank pages with signals that can be studied and manipulated, so a whole industry grew up around producing pages that rank rather than pages that help. The result is familiar: recipe posts with a life story above the ingredients and near identical articles answering the same question.
Generative AI poured fuel on this. It now costs almost nothing to produce thousands of plausible sounding pages, and low effort AI text has flooded search results, marketplaces and social feeds. At the same time, search engines started answering questions directly. A Pew Research Center study of Google searches made in March 2025 found that users clicked a regular result link on 8% of visits when an AI summary appeared, compared with 15% when it did not. Links inside the summary itself were clicked on just 1% of visits.
Fewer clicks means less revenue for the sites that produced the original information. Some respond by publishing more cheap content, some put up paywalls and some shut down, which leaves the open web thinner and the AI answers with less fresh material to draw from.
Paywalls, app pushes and enshittification
As ad money thinned out, many publishers moved to subscriptions. Paying for good journalism is reasonable, but readers now face a separate paywall, login and newsletter prompt on nearly every site, with no easy way to pay once for a little of everything.
Platforms add their own pressure, nudging you from the website to their app, where they get more data and more notifications.
Cory Doctorow coined a word for the bigger pattern: enshittification, which he expanded into a 2025 book. His argument is that platforms first treat users well to grow, then shift value to business customers such as advertisers and sellers, and finally squeeze both groups to pay shareholders. Search results crowded with ads, social feeds full of recommended posts instead of the accounts you follow, and marketplaces packed with sponsored listings all fit the description.
Ads sold against your profile, not the page you read.
Pages built to rank, now mass produced with AI.
AI answers keep visitors on the results page.
Services trade user experience for revenue.
How to get a better web in 2026
You cannot fix the economics of the internet alone, but you can change what reaches your screen.
- Use a content blocker. uBlock Origin remains the best known option in Firefox. Chrome dropped support for older Manifest V2 extensions in 2025, so Chrome users now rely on Manifest V3 blockers such as uBlock Origin Lite, which are lighter but less flexible.
- Switch to a privacy minded browser. Firefox, Brave and the DuckDuckGo browser block many third party trackers by default. Safari also limits cross site tracking out of the box.
- Bring back RSS. Many sites still publish feeds. A reader such as Feedly, Inoreader or NetNewsWire gives you a chronological list of new posts from sources you chose, with no algorithm deciding what you see.
- Try other search engines. DuckDuckGo, Brave Search and Startpage offer cleaner results pages and less tracking.
- Turn down the AI layer when you want sources. Adding a site name or using the Web filter in Google narrows results to classic links, which makes it easier to reach the original page.
- Pay for a few things you value. A couple of subscriptions to sources you trust, or newsletters from independent writers, do more for the web you like than any setting.
FAQ
What does enshittification mean?
It is Cory Doctorow’s term for the way online platforms decline: they start out good for users, then shift value to business customers, then squeeze everyone to maximize profit.
Why are there so many cookie banners?
Privacy laws require consent before many kinds of tracking, and most sites still rely on tracking based advertising. Instead of tracking less, many sites ask for permission on every visit.
Do AI search answers hurt websites?
Research on Google searches shows people click on fewer result links when an AI summary appears. That reduces traffic and ad revenue for the sites the answers are based on.
Is an ad blocker legal to use?
Yes, using a content blocker in your own browser is legal in the US. Some sites may ask you to disable it or offer a paid option instead.
The verdict
The web broke slowly, one profitable compromise at a time: tracking ads, search spam, consent friction, paywalls and now AI answers that skip the source. The open web is still there underneath. A good blocker, a privacy focused browser, an RSS reader and a few paid subscriptions bring most of it back into view.
Updated October 2026: rewritten as an explainer on why the web feels worse, including the Chrome cookie reversal, AI search answers and practical fixes.
Featured photo: Growtika on Unsplash.
About this article: GeekBlog covers U.S. technology news, AI, phones, smartwatches and gaming. Every story is written and checked under our Editorial Policy. Spotted a mistake or have a story tip? Contact our editors.
