The pattern recently is to collect every individual service degradation and present them as all equally significant.

Erase the severity and then present them all as “GitHub outages” or reduce it to an uptime graph.

I’m not happy with GitHub’s recent major outages either, but there is an ugly side of the pile-on where we’re getting these vibecoded attention seeking websites and social media posts to collect upvotes, likes, karma, and attention that blur the lines between small service degradations and total site outages to be more dramatic.

I think it is fair to blame Github if they repackage other services. We run a much smaller service than Github and have all sort of fallbacks to different providers and different models.

Depends on who's hosting the models.