I received my copy early and have played a couple of hours. It's a really impressive technical feat to play this in handheld mode, and I think the Joy-Con both offer good enough controls to make aiming work. I haven't had any trouble taking down groups of enemies competently, deftly jumping around and utilising melee attacks where necessary.
The game does indeed utilise the docked GPU clocks. Comparing two screenshots reveals a noticeable bump in resolution. I suspect when docked the game is either locked at 720p or is closer to 720p. The actual HUD isn't scaled beyond the same 720p found in handheld mode however, which is disappointing. Text is crisp on the Switch's display but blurry upscaled to a higher resolution TV.
Pixel counting one
handheld mode screenshot that I took, where stair-stepping around the edge of this one texture is obvious, netted 540p. But note that the game could be utilising dynamic resolution scaling. I haven't found enough easy edges to do any more reliable pixel counting.
It's impressive, and I find it plays well, but I also notice that when docked the game isn't really pushing the Switch. Putting my face next to the dock after ten minutes of play, the fans spin at a much quieter and slower rate than I was expecting. By comparison, Super Mario Odyssey pushes the Switch hard and the fans tend to spin at their maximum rate, also producing audible noise near the system.
I don't know. Either whatever he is using to record doubles the frames on the Switch version or this is shenanigans (or youtube can't run 30fps in 60fps mode and just double the frames).
I believe it's from the same YouTube channel that made hoax comparisons in the pass, for example for Snake Pass. Some suspect they just set the PC version to different settings for all the consoles.