Lots of good points in this thread, especially the insights from devs about the pain/gain tradeoff. I look forward to more native Apple Silicon apps, including games.
As I see it the M1 Macs offer some compelling attractions: the power efficiency of the hardware foremost.
Looking at the software, on the downside we lost easy Windows virtualisation and many applications incur a modest penalty by having to run via Rosetta 2. The upside is that native applications harness the performance efficiency of Apple Silicon. Part of the upside was meant to be that it would be easier for developers to leverage their work on iOS/iPadOS to bring more apps to the Mac. In theory, you could see the potential library of iOS/iPadOS ports marking up for some of the loss (e.g. Windows virtualisation).
Ultimately, it will take time to see how things pan out. As owner of two M1 Macs (MBA (8/512) and mini (16/1TB)), I fear that Apple is a victim of its own success with Rosetta 2. It works so well that there is little sense of urgency to get apps running native.
In short, faced between a choice of Mac app running via Rosetta 2 (optimised UI) or an iPad version of the same app running natively on an M1 (sub-optimal UI), the Mac version will almost always be the better choice right now.