Somewhere on your hard drive, or buried in an old email thread, there's a link you thought was permanent. Maybe it was a great article about net neutrality. Maybe it was a recipe. Maybe it was the source you cited in a report that your boss still references. Whatever it was, there's a solid chance that if you clicked it today, you'd be greeted by a 404 error, a GoDaddy parking page selling the domain for $1,499, or — worst of all — a completely different website that has quietly colonized the old URL and now sells ergonomic chair cushions.
Welcome to link rot: the slow, unglamorous decomposition of the web's connective tissue.
The Numbers Are Genuinely Alarming
This isn't a vague, hand-wavy problem. Researchers at Harvard Law School's Perma.cc project found that roughly 50% of URLs cited in Supreme Court opinions are no longer accessible. Supreme Court opinions. The kind of documents that are supposed to be permanent records of how this country interprets its own laws. If the highest court in the land can't keep its citations alive, your 2015 Tumblr post about artisanal hot sauce never stood a chance.
A 2021 study from Pew Research found that about 25% of all web pages that existed at some point between 2013 and 2023 are now gone. Gone as in: not archived, not redirected, not even acknowledged. Just absent. And that figure likely undersells the damage, because it doesn't account for pages that technically exist but have had their content swapped out entirely — the zombie link problem, where the URL lives but the original content has been replaced by something completely unrelated.
The internet, it turns out, has the institutional memory of a goldfish with a head injury.
How a Link Dies
Link rot doesn't happen all at once. It's a slow process with several distinct stages, each more depressing than the last.
First, someone publishes something online. A news article, a forum post, a product page, a government resource — it doesn't matter. It exists. People link to it. The link spreads.
Then, quietly and without ceremony, something changes. The website gets redesigned and the URL structure shifts. The company gets acquired and the old domain gets shut down. The hosting bill goes unpaid. The CMS gets upgraded and the slug gets reformatted. The writer leaves the publication and the publication decides to purge older content to save server costs. The startup runs out of money. The blogger gets bored and lets the domain lapse.
None of these events send you a notification. No one emails the people who linked to that page. The web has no forwarding address system. One day the link works; the next day it doesn't. And every document, article, or chat message that contained that link becomes slightly more useless.
Multiply this by billions of pages and a couple of decades, and you start to understand why the web feels increasingly like a city where half the buildings have been quietly demolished overnight.
The Archival Problem Nobody Wants to Fund
The Internet Archive — specifically its Wayback Machine — is one of the genuinely heroic institutions of the digital age. A nonprofit that has been crawling and preserving web pages since 1996, it holds over 800 billion web pages in its collection. It is, without exaggeration, one of the most important libraries ever built.
It also runs on donations and is perpetually underfunded.
In 2024, the Internet Archive was hit with a cyberattack and a series of copyright lawsuits from major publishers, temporarily taking it offline and threatening its long-term operations. The organization that exists to preserve the web was nearly destroyed by the web's own legal and security failures. The irony would be funny if it weren't so bleak.
Meanwhile, the platforms generating the content — the Facebooks, the Twitters-turned-Xs, the Reddits — have no meaningful commitment to preservation. Twitter's API changes effectively destroyed years of third-party archival tools. Reddit's own history is increasingly inaccessible as subreddits get banned or go private. These platforms generate enormous amounts of cultural content and then treat that content as a liability the moment it stops being profitable to host.
What This Actually Breaks
It's tempting to shrug at link rot as a cosmetic problem. Pages go away. So what? The internet is big.
But the consequences are real and compounding. Academic research loses its citations. Journalism loses its source material. Legal arguments lose their evidentiary links. Public health information — the kind that gets shared during a crisis — disappears and gets replaced by misinformation that has better SEO. History gets rewritten not through malice but through neglect.
And on a smaller, more personal scale: the things you saved, shared, and built digital memories around quietly stop existing. The article that changed your mind about something. The tutorial that helped you learn a skill. The forum thread where someone solved a problem you had. Gone.
The web promised permanence. It delivered a sandcastle.
The Fixes That Exist and Why Nobody Uses Them
There are technical solutions to link rot. Persistent identifiers like DOIs (Digital Object Identifiers) work reasonably well in academic publishing. Services like Perma.cc let journalists and lawyers create archived snapshots of URLs at the time of citation. The Wayback Machine's Save Page Now tool lets anyone create an instant archive of any public URL.
These tools exist. They work. And almost nobody outside of specialized professional contexts uses them, because they require deliberate effort and nobody has made them frictionless enough for casual use.
Browsers could display link health indicators. CMS platforms could automatically archive outbound links at the time of publication. Social platforms could preserve content instead of purging it. None of these things are technically impossible. They're just not profitable, so they don't happen.
Instead, we collectively agree to keep building on a foundation that erodes a little more every day, linking to things we hope will still exist tomorrow, and acting surprised when they don't.
A Web That Forgets Itself
There's a version of the internet that takes its own history seriously — that treats the preservation of information as infrastructure rather than an afterthought. We don't have that internet. We have this one: a web that generates content at extraordinary speed and discards it at nearly the same rate, leaving behind a trail of dead links and broken references that gets longer every year.
The next time you paste a URL into a document or a message, take a moment to appreciate it. It's alive right now. It's working. Someone is still paying to keep that server running, that domain registered, that content accessible.
Give it a few years.
The meter's running.