August 13, 2026
Link rot, but make it drama
Where did the old web go? We followed 657,607 links to find out
Nearly 8 in 10 old links are dead, and the comments turned it into a web funeral roast
TLDR: A revived backup of a tiny link-sharing site showed that nearly 8 in 10 old web pages from 2009 to 2014 are now gone or broken. Commenters turned it into a mix of obituary, nostalgia war, and joke fest over whether this was truly the “old web” at all.
The big number here is brutal: after chasing more than 657,000 old links from a resurrected Macedonian short-link site, the team found that 76.7% no longer led to a real loading page. In plain English: the internet lots of people remember and shared in the early 2010s has fallen apart, with small blogs, forums, photo pages, and local news hit far harder than giant platforms like YouTube and Google. But the real fireworks were in the reactions, where the mood swung from mourning to stand-up comedy in seconds.
One camp treated the whole thing like proof that the web was built on sand. One commenter flatly called disappearing pages one of the web’s biggest design mistakes, grieving lost game logs and vanished fan-made history. Another immediately turned the title into a generational fight: “09-14 is not even close to the old web for me,” they scoffed, basically saying, “kids, please.” That sparked the classic online debate over what counts as the “old internet” at all: Facebook-era links, or the prehistoric days of printed website directories.
Then came the comic relief. Someone was distracted by the resurrected project’s tiny, perfect domain name: “Wow, that is a great domain!” Another joked the site might have been overwhelmed by curious readers and asked, like a true internet archeologist, “Is there a torrent?” Underneath the laughs, there’s a real sting: this wasn’t just dead pages, but a snapshot of an entire online community — and a chunk of it is now a ghost town.
Key Points
- •A crawl of 655,178 safe historical links restored from the 0.mk URL shortener found that 76.7% no longer returned a loading page in August 2026.
- •After removing duplicate destinations, 492,620 distinct crawlable URLs remained, and 78.7% of them still did not load.
- •At the unique-URL level, 55.0% failed at the network layer and 23.7% returned HTTP errors; 404 was the most common result, while 403 and 429 may partly reflect crawler blocking.
- •Of 133,605 crawlable hostnames, only 34,827 had at least one URL load, while 98,778 had no loading URLs at all.
- •The article finds that large centralized platforms such as YouTube, Wikipedia, and Google properties survived better than personal blogs, forums, local news sites, and photo hosts, and it documents losses in Macedonia's web history through now-closed media outlets.