The Web Has Amnesia: Why Half the Links You Click Lead Absolutely Nowhere

Picture a library where, every few years, a significant portion of the books simply vanish. Not stolen, not checked out — just gone. No record of their existence. No forwarding address. The shelves quietly close the gap, and nobody files a missing persons report. You'd call that a catastrophe. On the internet, we call it Tuesday.

Link rot — the process by which URLs decay into 404 errors and digital content evaporates without ceremony — is one of the web's most persistent and least discussed disasters. It doesn't make headlines because it happens slowly, one dead link at a time. But the cumulative damage is extraordinary, and the reasons it keeps happening reveal something genuinely troubling about how little we value the permanence of online information.

The Numbers Are Ugly

Researchers at Harvard Law School's Perma.cc project found that approximately 50% of URLs cited in Supreme Court opinions no longer work. Supreme Court opinions. Legal documents meant to anchor the interpretation of law for generations, footnoted with links that now return blank pages or domain-squatter ads for insurance quotes.

A 2021 study from the Pew Research Center found that roughly a quarter of all web pages that existed between 2013 and 2023 are no longer accessible. Not archived somewhere obscure — just gone. The Wayback Machine at the Internet Archive has done heroic work preserving what it can, but it's a single nonprofit organization playing digital janitor for the entire internet, and it can only catch what it crawls before the lights go out.

For context, the average lifespan of a web page before it disappears or changes significantly is estimated at roughly 100 days. The Roman Colosseum has been standing for nearly 2,000 years. Your favorite article from 2009 about the best apps for your first iPhone has a worse survival rate than a mayfly.

The Many Ways a Link Dies

Link rot isn't one problem — it's a family of problems, each with its own flavor of dysfunction.

The most common culprit is simple neglect. Personal blogs, small publication sites, and independent projects get abandoned when their creators move on, lose interest, or can't justify paying hosting fees for content nobody's reading anymore. The domain expires. The registrar reclaims it. Sometimes it gets snapped up by a domain squatter who replaces years of thoughtful writing with a parking page full of ads. The original content doesn't go anywhere dramatic. It just stops existing.

Then there's the corporate deletion — more aggressive, more damaging, and somehow more offensive. In 2019, MySpace announced it had "accidentally" lost 12 years' worth of user-uploaded music during a server migration. Approximately 50 million songs from 14 million artists, gone. The quotes around "accidentally" aren't scare quotes — that's genuinely how they described it. An accident. Twelve years of cultural output, uploaded by people who trusted a platform to hold it, wiped out with the casual indifference of someone clearing a desk.

Media companies do this too, with less fanfare but equal damage. Gawker's archives were nuked following the site's legal collapse. Various regional newspapers have quietly deleted years of local reporting as they've consolidated, merged, or folded entirely. Quartz deleted its entire archive in 2023 as part of a "strategic reset." The journalism didn't stop being valuable. The company stopped caring about it.

The URL Was Supposed to Fix This

Here's the irony: the web's architecture was explicitly designed with permanence in mind. Tim Berners-Lee's original vision for URLs — Uniform Resource Locators — was that they would function as stable, permanent addresses for information. He even wrote a famous essay in 1998 titled "Cool URIs Don't Change," arguing that a good web address should work indefinitely, that changing URLs was a failure of stewardship.

That essay is still online, which feels like either a hopeful sign or an elaborate joke depending on your mood.

The problem isn't technical — it's economic and cultural. Maintaining URLs requires ongoing effort: updating server configurations during platform migrations, setting up redirects when content moves, actually caring about what happens to old pages when you redesign your site. None of these things generate revenue. None of them show up in quarterly metrics. So they don't get done.

Content management systems make it easy to change URL structures with a few clicks and nearly impossible to automatically redirect every old link. Platforms get acquired and their URL schemes change overnight. Publications restructure their archives and shrug at the broken links left in their wake. The incentive to maintain the past simply doesn't exist in an attention economy obsessed with the present.

The Institutional Memory We're Burning

What's actually being lost here isn't just convenience — it's the connective tissue of public knowledge. Academic papers cite sources that no longer exist. Journalism references background reporting that's been deleted. Wikipedia editors fight constant battles to replace dead citation links with archived versions before articles degrade into uncited assertions.

Local news is particularly brutal in this regard. The historical record of American communities — the school board meetings, the zoning disputes, the small-business profiles, the obituaries — lived primarily in the archives of local newspapers. Those papers have been closing at a rate of about two per week for over a decade. When they go, their archives frequently go with them. The communities they covered lose their recorded history as casually as losing a receipt.

The Archive Isn't Enough

The Internet Archive is genuinely one of the most important institutions in existence, and it operates on a budget that would make a mid-sized tech startup laugh. Brewster Kahle built something that functions as humanity's backup drive for the open web, and it runs on donations while Google runs on hundreds of billions in revenue.

The disparity is clarifying. The tech industry has made enormous amounts of money from the web's content ecosystem while investing almost nothing in its preservation. Archiving is a cost center, not a profit center, so it gets outsourced to a nonprofit and mostly ignored.

There are better models. Perma.cc lets legal and academic authors create permanent, verified snapshots of cited URLs. Some academic publishers use DOIs — Digital Object Identifiers — that persist even when URLs change. The tools exist. The will to use them, outside of narrow institutional contexts, largely doesn't.

You Can't Link to What's Gone

The web's power was always the link — the ability to point, to reference, to build on what came before. Every dead URL is a severed connection in that network, a hole in the argument, a citation that proves nothing because the evidence has vanished.

We've built the greatest information infrastructure in human history on a foundation that treats permanence as optional. And we're surprised, every single time, when we reach for something and find nothing there.

The 404 error isn't a technical message. It's a eulogy. We just stopped reading them.