EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Aug 13·5 min read
Where did the old web go? We followed 657,607 links to find out
TL;DR
A restoration of a 0.mk database revealed that 76.7% of 657,607 links from 2009-2014 are now dead, highlighting the web's rapid decay.
✦ Why It Matters
Consider archiving important web content to prevent loss in the future.
Key Takeaways
- Assess the longevity of links in your projects.
- Implement archiving strategies for important web content.
- Monitor error rates for your URLs regularly.
- Consider using centralized platforms for better content stability.
Full Summary
The 0.mk URL shortener, popular in Macedonia from 2009 to 2014, had its database restored, allowing researchers to analyze 657,607 links. Following these links in August 2026, they found that 76.7% of the crawlable links did not load, indicating a high rate of web decay.
The study revealed that only 21.3% of unique URLs loaded successfully, with many returning errors like 404 (not found) or 403 (forbidden). The analysis also showed that larger, centralized platforms like YouTube and Wikipedia fared better than smaller, personal sites.
This data serves as a historical snapshot of a community's online presence, illustrating the fragility of web content over time and the challenges of preserving digital history.