I think people like to take the easy way out of declaring those with a different mindset "evil." Everyone is the hero of their own story, and honestly there are multiple incompatible-but-internally-consistent models of how technologies can and should work. I think it's more useful to recognize these things than to write off a competing mindset (especially when the competing mindset is in a position of power).
Consider incentives from Google's standpoint. They want to provide users a safe and secure experience. They want to simplify maintenance of software and provide developers the ability to simplify maintenance of software (a problem simplified by chopping the unbounded set of possible user agents down to a blessed, vetted subset). They have the resources to make their site screen-reader compatible, so they're not concerned about damage that could be done to screen-readers because they'll just bless one and support it. And, of course, they implicitly trust themselves to do all this.
In that ecosystem, Weiss's viewpoint is completely reasonable. The old model of the web is old, and led to gestures broadly at all the bad things about the web today... fraud, users getting owned, CP, botnets, misinformation factories. I can definitely see the viewpoint where someone concludes "It's time for a new model, and this company has the resources to do it."
I don't agree with him (and in fact I think the idea will fail; I think Google actually overestimates its ability to provide an equivalently-good user experience to what we have now if they aren't leveraging the unpaid labor of other vendors putting the effort into making their own houses work with Google's house without Google even being aware of their work). But I think it's useful to wrap our heads around how one gets into that headspace without thinking oneself a monster.
As they say: "the road to hell is paved with good intentions". Wanting to fix the world by taking complete control of it is one of the most trivial examples of a plan that should be immediately labeled "evil", as, if nothing else, "absolute power corrupts absolutely".
This plan doesn't take complete control. It provides a mechanism for a web site to delegate trust on UA configuration authenticity to a third party, or even to itself via side-channel.
Nothing in the proposal requires the third party be Google. The proposal does decrease the control the user has over their own hardware, in the sense that it provides a channel for a site to decide the user-agent / hardware stack is the wrong pedigree to serve; that's not universally considered evil either (few people really get bent out of shape that you need a Nintendo Switch to use Nintendo Switch Online services).
Google sells ads. They want to kill ad blockers. This is how.
> Weiss's viewpoint is completely reasonable
Chasing diversions around in circles is not neutral. Someone wins by default. Diversions exist and they exist to tempt you into poor attention allocation decisions. This is not about safety, security, and providing an excellent experience. It's about ads and making sure you can't stop them.
It's extremely likely it's about both. It can be both about making it hard to skip ads on YouTube and about making it hard for somebody to replace human users with automated devices.
