We're in the typical start-stop-resume release process Tesla has used for all of the previous releases.
Tesla's internal testing and their hand-picked beta testers either miss major problems with the release - or Tesla decides to ship the software even with these major bugs.
They start distributing the software on a seemingly random basis to a small subset of cars, and then trickle the releases out to more cars on a daily basis - until - a major problem is encountered. The distribution then appears to slow down considerably or stop entirely - for days or weeks, and then they start the process over when they believe they've fixed the problems.
And when that release (which gets even less testing than the initial major release) is found to have major problems - the process is repeated, until eventually we get a relatively clean release.
It doesn't have to be this way. Other companies have found ways to both deliver on aggressive schedules AND with acceptable quality. But that may require some major changes to Tesla's software development process (more planning, more reviews of planned features, open beta testing, ...), and, at least so far, Tesla evidently isn't ready to admit they have this problem - and prefers to continue using the same flawed development process - with the same results for each major firmware release.