It's only marginally better. So far nothing has been better than AutoCAD's approach, for me - all GUI commands are echo'd into the command-line as you do them, so you can see what commands you're using if you want to key them in later. Negligible hotkeys because the commands work so well and are more memorable than hotkeys.

Google proved many years ago that average users are more than happy to use a search bar as their primary user interface to everything. I've seen so many users that to go to any website never use bookmarks, addresses, shortcuts, anything, just type it vaguely into a search bar, and that's been some of their norm for a decade or more.

It makes sense for Microsoft to front-and-center the search box because that's what users use the most. That's what Google has been doing for its entire existence. It's about as dumb and simple a UX as possible.

Beyond that it's a "Why not both?" situation. A search bar isn't very discoverable, but it is easy and typically "does what the user wants", eventually, depending on how much effort goes into the "man behind the curtain". But you can have search and try for a discoverable UX at the same time. The Settings app many deride is something like Attempt #15 at making Windows settings discoverable. For better or worse, with as many settings as Windows has, making that discoverable is a herculean task, if not a sisyphean task. Every attempt has annoyed some people. The Control Panel has always been painful to use. The Settings app tried for a somewhat clean break and of course there are ton of opinions on it, because it moved cheese and its so different (though is it really?) from the Windows 3.1 Program Manager folder some people seem to expect still frozen in perfect amber from when they first learned to use Windows... There's no pleasing everyone, and there's no perfect path to discoverability or usability.

There's no "smart" path on either side; one requires currently unimaginable tools to read people's minds based on tiny text fragments they through into the void of a search box, and the other to anticipate every users needs and somehow make them all discoverable exactly when the user needs them. (One requires telepathy/telempath and the other prophecy, perhaps.)

I mentioned a "happy path", but that's extremely subjective. (One user loves it if search works great; another if they know just what to click; a third if they have a good CLI to automate it; etc.) It's also clear that there are many more fewer happy paths, than paths in general, and nearly impossible to "pave" all the really good paths for people.