I still think there's room for both types of processor. Apple silicon is the ultimate evolution of Steve Jobs' "performance per watt" metric, which at the time was a huge reason for the Intel transition. It's one of the same reasons they're citing for the ARM transition.
But raw performance, or even raw performance per watt, isn't the only metric that matters. As far as I've read, Apple silicon still limits users to a single external display. A lot of software is still not ARM compatible (and, just like with the PPC transition, some will never be) - I'm not sure how long Rosetta 2 will hang around, but there's still plenty of things depending on the translation layer. Of course, the simple fact that running Windows on M1 is still a hodgepodge of "maybe things will work, maybe they won't", and at least I personally know of many Mac users who heavily depend on a Windows VM for important tasks.
If you seek pure raw performance and/or maximum battery life, M1 is a win. But for all the talk of how fast M1 is, we mustn't forget the challenges users face with practical day-to-day usage. The fastest chip in the world means nothing to a user if the software that user needs to run either doesn't run at all or runs through a translation layer which makes it slower than a native Intel CPU.
The Intel transition at least brought Boot Camp with it, allowing native x86 OSes to boot on the bare hardware. I'd argue the Intel transition actually made Macs even more attractive, because now you didn't have to choose - you could dual-boot Windows on your Mac, or shortly thereafter, boot a VM with near-native performance. Windows on ARM stil has a way to go before it'll be truly on par and competitive with x86-64-based Windows. The same applies to M1.