This makes me wonder: what actually changed?
If I'm resisting the temptation to use these kind of interviews for buzzword bingo and focusing on what appears to be said between the lines:
Since the post is from July, neither the AI slop generator, the open-ended slop questions nor the 1 rep chat for everyone experiments were live yet.
That leaves us with Discussions and the two 1 rep lobby chats. None of which actually means a change to how the site is moderated, apart from the company abandoning these two experiments pretty much instantly, leaving them with no moderation at all. As in: if it ain't moderated, it ain't over-zealously moderated.
Which in practice meant opening up these channels as a free playground for spammers, trolls and low-life "code dumpers" who think every text box on the Internet is an AI prompt etc etc. If not for a few brave SO diamond mods who did moderate all this Discussions/chat slop, even though the company was lying to them all the time and said that there would be staff support for these features.
Since then we've had Discussions abandoned, but just when we thought that surely it wouldn't be possible to design a site feature worse than Discussions... enter open-ended questions.
And then we have 1 rep for any chat, which had very little impact on SO.
Meanwhile in SO Q&A: over-zealous moderation and general toxic behavior is still thriving, as it always did, thanks to the flawed core site mechanics. Notably however, there never existed any programming forum throughout history that didn't have problems with toxic behavior, so it is as much an unsolved problem now as it was before we had SO and Reddit.
What seems pretty certain to me is that none of the peripheral feature experiments we've seen in the past year will save SO. Much too little, much too late, far too sloppily implemented.