The unexplored outcome of hybrid is that there's no such thing as hybrid.
In a hybrid model, you come into the office and find people sitting in chairs with a headphone on. Because they're just as much in calls as they were at home. Because part of the meeting participants on any given day is not there.
Hence, this nice physical face-to-face meeting, a supposed benefit of returning to the office, does not materialize.
Well, perhaps you'll at least do some ad-hoc brainstorming with your colleague at the coffee machine. Sure, but you need to put the outcome of that in writing...for the people that are not there.
Hence, this mode of virtual working cannot be undone. You can't do a "little" virtual working. When you come into the office you largely continue to work as if from home. Perhaps without the sweatpants.
For the record, I'm not anti-office. I enjoy the socializing, lunch walks, etc. And in our increasingly touchless home-based society, it's not that bad of a deal to get out of the house for 2 or 3 days a week. Finally, we should empathize with juniors. In my experience, remote mentoring is a piss poor substitute for the real thing.
It does in my experience. 1 on 1s and small meetings often do have all of the attendees present, especially when in-office days are designated. Some meetings do have people attending virtually, but a good conference room setup with large screens and good interfaces for conferencing make it more comfortable than being stuck in front of your computer. Regardless, you are in the room with many people face-to-face, and you can discuss things with them after the meeting and build relationships.
I see your point, because sometimes I have to find a room in the office just to have a 1:1 virtually, which arguably is more of a hassle than just taking the call at home. But I do find that hybrid works well overall in my experience, and the interpersonal benefits are large even though not everyone is in the office at the same time.
Since when does this have anything to do with AI? Commercial/enterprise software has always been this way. If it's not going to cost the company in some measurable way issues can get ignored for years. This kind of stuff was occurring before the internet exists. It boomed with the massive growth of personal computers. It continues to today.
GenAI has almost nothing to do with it.