Silicon Valley Technology Commentary & Archives · Est. 2006 3,037 Posts · 2006–2026

October 30, 2026

October 30, 2026 · 4 MIN READ · BY LOUIS GRAY

Our Imperfect Decaying Web Desires Permanence

Our Imperfect Decaying Web Desires Permanence

Broken hyperlink bridges alongside enduring pathways leading to a lit archive beacon [Image via Gemini]
Broken links vs. paths built to endure. [Image via Gemini]

On the web, there's not much more sad than a broken link. A fundamental promise of the web, and a defining characteristic of it, has always been that you can get from one point to another, with blue underlines guiding you to the next discovery. A link that fails is a broken contract between the source and the destination. The initial author who sent you the link promised you value, and the target address didn't pay out. It's frustrating for everyone. But this expectation of permanence, that once a link is published, it works forever, has proven to be a pipe dream. 

The Pew Research Center, in a 2024 study, found that nearly 40% of links from just a decade prior had stopped working, and fully 20% of links on the web failed within two years. An earlier 2014 Harvard study said fully half of links cited in Supreme Court opinions no longer went anywhere. While we've thought we were building a foundation of discovery, content is flowing out the bottom of the sieve practically as quickly as we can fill it, thanks to ephemeral design, social content perceived as lower value, and countless services that come and go.

Until recently, I was guilty of this, with my own blog being shuttered. I assure you, being a champion of the web, it felt terrible. But even when it would have been easier to start from scratch with another domain or platform, I held to an implicit contract with my past self and untold readers that I could only turn the lights back on when all my posts had been restored, and whatever links had been built in my direction would work again. Working with Antigravity and Cloudflare, I rebooted the blog, confirmed domain resolution, inherited the proper link structure, and even reattached old blog comments from more than a decade ago, due to my commitment to the bit. It's how the web is supposed to work.

But what about all those links I had curated, by hand, in my more than 3,000 existing posts? Reader, I have some bad news.

Screenshot_20261030-232732.png
Despite best efforts, thousands of links I made are gone forever.

I asked Antigravity to do a link audit and report of the tens of thousands of links I'd made over the years -- to other blogs, to companies, social profiles, news and all manner of destinations. Of the nearly 30,000 links, coming to under ten links per post, just over one third are dead, taking people on a one way ticket to nowhere. You can see that report here: https://blog.louisgray.com/linkaudit. 

So what happened, aside from good intentions and what not? 

Screenshot_20261030-233701.png
I don't know what I expected.

Unsurprisingly, the social stars of a bygone era are a big contributor to my amassing thousands of trips to nowhere. Almost 1,000 links to Friendfeed didn't do me much good after Facebook gobbled it up and smothered it with a pillow. Many of us walked away from Twitter (due to its toxicity or other reasons) and what was supposed to be a safe place to link ended up being a transitory target. Even Google, be it Reader or Buzz or what have you, hasn't been without its flaws. And that's not even getting started on all those Web 2.0 startups that didn't win, despite best effort.

Back in 2007, we had one of our first major blog skirmishes when I said "Internal linking on some tech blogs is out of control", as outlets used keywords to point to themselves instead of their topics du jour. But there's clearly value in it! If you can link to your own posts, at least you control the destiny, and it's up to you whether they still live -- like my 2013 post talking about "Tech's halflife and accelerating forgetfulness". As I said then, "History can be dull, yes. But it can also be fascinating." My hope would be that we build, together, a web that expects links to work and search engine (or AI queries) to return data that can be relied on to still work, not just months down the line but years.

So much of the web has gone dark, behind paywalls and login gates, that the content we push into the open web, aimed to enrich or discuss, inform and discover, should be designed for permanence. We're all very lucky sites like Wikipedia and the Internet Archive exist, to document and backup what we've built, but every time I try to visit an old blog that used to be a mainstay, or I see an RSS feed I long ago subscribed to just not resolve, it's another gap in our story -- another dry valve that used to flow. Let's try to fulfill this social contract, that when we build, we build for domains that last, paths that endure and content that stays up.