Based on my experience (nearing 30 years of system administration), one thing led to the others. They lost the capacity to handle lots of simultaneous requests when Google cut them off, so they pushed the change to rate limit at API level, but since they don’t seem to have QA or testing anymore, this led to the front-end hitting the limit, but the front-end didn’t have an error handler for this situation, so it just kept trying.

Tl;Dr, Twitter 2.0 is a shit show.

It’s actually bellow amateur hour. The best metaphor I can make for how things are working there right now is a bi-polar Chihuahua on crack having an epileptic seizures.