OpenAI scraping Reddit through redlib instances

6 months ago (hcrypt.net)

The author used to run a public Arch mirror under mirror.ext4.xyz, so it's not exactly an unknown domain.

Combined with the fact that a lot of their self-hosted stuff, including the Reddit front-ends, are in the Certificate Transparency logs [1], it's not hugely surprising that web crawlers would run into them.

[1]: https://crt.sh/?q=ext4.xyz

  • I am definitely not surprised. It's quite normal, as I stated. What I was trying to say was they are abusing the private frontends to get around legal restrictions of aggressively scraping the web