My 2¢: it's cool, but it doesn't work nearly as well as described here, and this is definitely something that deserves the "beta" label. What's mystifying to me is why this is difficult, because I wrote my own version of this to work between Macs and it works far better than Apple's version, at least theoretically (as in, it works for me, and isn't broken in the ways that Apple's is broken, it's just not ready for release because I haven't polished the UI yet).
Apple's implementation requires you to have all the devices signed into the same iCloud account…well, guess which two computers on my desk I would like to share my input devices across? Yeah, my personal and work laptops, but I'm definitely not going to sign into iCloud on one of them, so the feature is useless for that right out of the gate. (And this is why I made my own thing, to be clear.) Putting that usecase aside, there's a Mac mini under the desk that I use for kernel debugging sometimes, and an iPad next to the monitor. Does it work with those?
For the iPad, if it's on, then sure, it works most of the time. Helpful when I am working on Swift Playgrounds I guess, since the Mac app doesn't have any of the app-making features I want to play with. If I didn't have the iOS Slack and Discord apps running on my Mac, I can see it being useful for that too. But otherwise, I don't really have a need to actually do anything on my iPad if I'm at my Mac, although I would be super interested to hear what people are using this for.
For the Mac mini, things are substantially worse. It's sleeping most of the time, because I'm not using it, and that just means Universal Control doesn't work. It's kind of hilarious because often I will poke it with my software, which uses functionality that Apple already ships (try screensharing into a sleeping Mac!) to wake up the computer, but Universal Control doesn't seem to use it. Even when the computer is on getting it to recognize the device is hit-or-miss, even though they're on the same home network and feet from each other. It seems worse at discovering the device than AirDrop is, which is saying something.
Speaking of which, even when you've got all your devices all linked up, the display configuration part is super janky, for reasons I think are extremely Apple. It looks super cool to do a keynote demo where the mouse approaches the screen edge and jumps over to the device on that side, but in the real world, people have computer layouts that don't match that. My monitors don't always match where devices are in the real world, so I frequently have the cursor appear on the wrong side of the display, and then I need to go into System Preferences anyways to fix the thing, so the magic is completely gone at that point. Plus, the configuration doesn't save and the preference pane to rearrange displays clearly didn't get the necessary love to do screen layout correctly (it's non-trivial, but not actually that hard…a fun exercise in geometric algorithms to make sure all the displays are contiguous and things like snapping work). So things will just jump around, and you'll find that certain displays can't be placed in certain positions. If you get things wrong you might lose your cursor in the "abyss", although I think it tries really hard to keep the cursor on a screen visible.
Honestly, I don't get why the feature is like this. Apple spent months working on this past the initial ship date, and it's still in "beta", and it still seems really finicky. I get the whole "things are hard actually and it's not always obvious why" but I made this thing (and I'd really love to share it, but I just can't yet) and I am not seeing why what Apple did is significantly harder than forming a secure network connection and forwarding Quartz events over it, which is what I'm doing, and seemingly works far better than their implementation does for the purposes of controlling another Mac. What am I missing?
Apple's implementation requires you to have all the devices signed into the same iCloud account…well, guess which two computers on my desk I would like to share my input devices across? Yeah, my personal and work laptops, but I'm definitely not going to sign into iCloud on one of them, so the feature is useless for that right out of the gate. (And this is why I made my own thing, to be clear.) Putting that usecase aside, there's a Mac mini under the desk that I use for kernel debugging sometimes, and an iPad next to the monitor. Does it work with those?
For the iPad, if it's on, then sure, it works most of the time. Helpful when I am working on Swift Playgrounds I guess, since the Mac app doesn't have any of the app-making features I want to play with. If I didn't have the iOS Slack and Discord apps running on my Mac, I can see it being useful for that too. But otherwise, I don't really have a need to actually do anything on my iPad if I'm at my Mac, although I would be super interested to hear what people are using this for.
For the Mac mini, things are substantially worse. It's sleeping most of the time, because I'm not using it, and that just means Universal Control doesn't work. It's kind of hilarious because often I will poke it with my software, which uses functionality that Apple already ships (try screensharing into a sleeping Mac!) to wake up the computer, but Universal Control doesn't seem to use it. Even when the computer is on getting it to recognize the device is hit-or-miss, even though they're on the same home network and feet from each other. It seems worse at discovering the device than AirDrop is, which is saying something.
Speaking of which, even when you've got all your devices all linked up, the display configuration part is super janky, for reasons I think are extremely Apple. It looks super cool to do a keynote demo where the mouse approaches the screen edge and jumps over to the device on that side, but in the real world, people have computer layouts that don't match that. My monitors don't always match where devices are in the real world, so I frequently have the cursor appear on the wrong side of the display, and then I need to go into System Preferences anyways to fix the thing, so the magic is completely gone at that point. Plus, the configuration doesn't save and the preference pane to rearrange displays clearly didn't get the necessary love to do screen layout correctly (it's non-trivial, but not actually that hard…a fun exercise in geometric algorithms to make sure all the displays are contiguous and things like snapping work). So things will just jump around, and you'll find that certain displays can't be placed in certain positions. If you get things wrong you might lose your cursor in the "abyss", although I think it tries really hard to keep the cursor on a screen visible.
Honestly, I don't get why the feature is like this. Apple spent months working on this past the initial ship date, and it's still in "beta", and it still seems really finicky. I get the whole "things are hard actually and it's not always obvious why" but I made this thing (and I'd really love to share it, but I just can't yet) and I am not seeing why what Apple did is significantly harder than forming a secure network connection and forwarding Quartz events over it, which is what I'm doing, and seemingly works far better than their implementation does for the purposes of controlling another Mac. What am I missing?