“The serious costs and widespread performance issues and outages caused by reckless scrapers has been on everyone’s mind in the sysadmin community as of late, and has been the subject of much discussion online.”
SourceHut’s terms now ban using its users’ data to train AI, and it put a proof-of-work wall in front of its services. Its robots.txt threatens to block violators’ whole subnets. GitHub is named as the host that sided with its own AI products. SourceHut is paid by its users, not by selling them.