It's in no way worse than running a single browser extension with overly broad privileges. If this method of adblocking gains more traction (quite possible, as Google keeps moving to damage in-browser ad blockers), I expect the implementation to receive a lot more scrutiny.
I think we have to face the reality that web browsers might no longer be considered "user" agents.
> I think we have to face the reality that web browsers might no longer be considered "user" agents.
I think too many techies are, much like yourself, contributing to the problem by refusing to move away from Chrome for "reasons"[1], and then compounding it by refusing to acknowledge that "web browsers" != "Google Chrome".
[1] The "reasons" are of dubious quality. Myself and many others are able to do all normal web-browsing from firefox or firefox forks with no functional or performance degradation.
> I think too many techies are, much like yourself, contributing to the problem by refusing to move away from Chrome [...]
Contrary to your assumptions, I've been quite vocal against the Chrome/Blink monoculture for a while. Unfortunately there is a legit case for it; several generations of "low-end" devices (anything older than 10 years basically), that are still quite capable and in common use, where the difference in performance between Firefox and Chromium becomes quite noticeable, especially as you try to watch video.
I don't think the problem is "techies", we have zero influence outside our own circles - see the historical rates of Linux adoption. The problem is we need the good ol' hammer of antitrust to start swinging again. We also need the regulators to be smart; if we get really unlucky, they will target iOS Safari instead. (This would be good in a healthy ecosystem, but would only serve to further entrench Google's position in the current situation.)
By the way, using a filtering/rewriting proxy has other merits, especially on said older hardware; you can rewrite the entire web page to make it more lightweight and accessible. Check out miniwebproxy[1] and medium-rare[2]. It's also quite simple to write one; you need maybe a hundred lines of Go to start getting results. I've been experimenting with integrating Readability[3][4]; and I think there's more potential to this approach.
> I think we have to face the reality that web browsers might no longer be considered "user" agents.
If by "web browsers" you mean specifically Chrome, yes. Firefox, Brave, and others are all committed to supporting MV2, and will continue to serve my interests as a user for the foreseeable future.
I think it is a mistake for programs that have SSL traffic to not have an option for a user-defined non-secure proxy (usually this would be for a proxy running on the local computer, rather than a remote proxy). A non-secure proxy would save energy (since then it doesn't need to decrypt and encrypt it twice) as well as allowing use of newer (or older) cipher methods in programs that do not support them.