Not to mention that until very recently even MS Teams sent you to a different product when you wanted to stream to 500 people. Even if it's now integrated, it's still a different product inside (and e.g. you could for example open a new window when you were in a 500 people "meeting" at the time when you still could not do so for a regular meeting).

You say "just more comfortable" but if you have two streams and one of them is on a channel you know to be unreliable (Jitsi) it's pretty guaranteed the unreliable stream is going to be down a significant percentage of the time. If you're a company with 500 people this isn't a comfort question, you're wasting probably hundreds of hours of your employees' time.

I think we're not on the same page about Jitsi being unreliable. In fact, it has been more reliable for me than Zoom in the past. Maybe due to the fact that I'm running Linux, I don't know, I haven't tried either on Windows.

For the corporate or training use case, this is not a problem. If you are worried about US law agencies, you shouldn't be using any system that isn't rooted in face to face communication for anything sensitive. (And even that is suspect with as small as bugged devices are today.)

There is a huge difference between requesting data that has already been collected and requesting Zoom/Microsoft/Google to record future data. The latter probably requires some serious intent. And of course, if I would want to be entirely safe from US law enforcement espionage then I would need to not use computers but whose use case is that?