He's made a strong case for what the original vision was, and maybe from the perspective of "social contract" that should be kept regardless of whether it is viewed as the "best thing" for Bitcoin. Obviously opinions strongly differ on the latter, and maybe that has to do with the context of us now living in a post-Snowden era which didn't exist when that "original vision" was defined. But the argument against the original vision being nodes all in data centers, most people using SPV, etc. is getting very thin.
There are security challenges with a network consisting mostly of light nodes.
Why aren't more people talking about ways to address them rather than using their existence as an excuse to prevent progress?
Largely because at the moment too many people are obsessed with the block size debate as a gladiatorial contest.
I really doubt that anyone is opposed to improvements to SPV security, although it is also the case (as we've seen with the block size debate) that what one person claims is an "improvement" may not be viewed by others as such.
But to the extent your point is valid, maybe people should take a step back from the block size, work on the other pieces, and then raise the block size again when the infrastructure is there to better support it. I've made that same point with respect to DoS protection and tragedy-of-the-commons issues. You don't just take away the limit, you address the issues that made it necessary, and
then it makes sense to loosen it or remove it.