So I did something dumb with my Saturday. I decided to go back through every post I wrote on here between 2011 and about 2015 and click every single outbound link, because I'd gotten it into my head that I wanted to fix them all before this blog hits its fifteenth birthday this November. I did not finish. I did not even get through 2013. But I learned some things and I'm mildly furious about a few of them, so here we go.
First, the scale of it. I went through 41 posts and counted 187 links. Of those, 61 are just gone — not redirected, not archived, actual 404s or the domain doesn't resolve at all. Another 34 redirect somewhere that has nothing to do with the original content, which is its own special kind of insulting. I linked to a great teardown of the original Kindle Fire back in December 2011 and that URL now points to a mattress affiliate page. A mattress. There is no version of events where that domain owner intended for tech blog traffic circa 2011 to end up buying a mattress in 2026, but here we are, and honestly the SEO logic of it kind of works, which makes it worse somehow.
The stuff that's genuinely gone, not just repurposed, is the part that actually gets me. Rdio's blog. Readability's whole site (the read-it-later app, not the concept). A handful of personal blogs from people who I'm assuming just stopped paying for hosting at some point and nobody noticed until I clicked a link in 2026 wondering what happened to them. One of them, a guy who used to write really sharp stuff about ARM chip design, I have no way of finding out if he's even still writing anywhere. His domain is parked. That's it. Fourteen years of somebody's writing, and the only trace left is however many people bothered to save it somewhere else, which based on my very unscientific Wayback Machine spot-checking is maybe 60% of it.
I want to be clear I'm not writing this to be nostalgic about Google Reader again, everybody who was blogging in 2013 has done that post already, myself included probably twice. This is more specific than that. It's the sheer arbitrariness of what survives. A throwaway comment I left on a forum in 2012 is still indexed and searchable. A 3,000 word post from someone who clearly put real work into it just isn't there anymore because their credit card expired and nobody around them thought to mirror it. There's no logic to it. It's not that the good stuff survives and the bad stuff doesn't. Its just whoever kept paying the bill.
The fix I've landed on, at least going forward, is boring but I think it's the right call: I'm going to start archiving anything I link to through the Wayback Machine's save-page-now tool at the time I publish, instead of trusting that the link will still resolve to something sane in ten years. It takes maybe eight extra seconds per link. I should've been doing this the whole time and I wasn't, mostly out of laziness, partly because in 2012 it genuinely didn't occur to me that half the internet I was reading would evaporate within a decade.
For the old stuff, I'm not going to fix all 187 links, that's just not a realistic weekend project and I have a day job. What I am doing is going through and swapping the fully-dead ones for archive.org snapshots where one exists, and just leaving a little strikethrough note on the ones that don't ("link no longer works, couldn't find an archive, sorry") rather than pretending the reference still means something. I got through 2011 and 2012 doing this. 2013 onward is still sitting in a text file on my desktop mocking me.
Anyway. My cat sat on the keyboard twice during this and once managed to actually publish a draft post that was just the string "asdfjkl asdf" as the title, which I caught about four minutes later, so if you saw that in your feed reader, that was her, not me, and no, she does not have opinions on link rot, she was just annoyed I hadn't fed her yet.
If anyone reading this has a mirror of that ARM blog I mentioned, or knows what happened to the guy who wrote it, genuinely, email me. It's bugging me more than it should.