Personally I hope firefox dies as fast as possible so we see some focus on good alternatives.
Gecko is not a good platform, there is a reason why people who use geckoview eventually all migrate away from it, the most recent example I can think of is wolvic, which hasn't replaced geckoview yet, but does have the version 1.0 of a chromium release now.
The sooner we get real alternatives to chromium and stop pretending that gecko is one the better. Currently servo is progressing really fast, has good APIs and usability for both a full desktop browser and embedded usecases (but still very immature).
The numbers you have quoted so far don't make a dent in the 400M though - we haven't even reached 1% yet. How much do you think Mozilla is spending on Firefox? How much of that is "extra" per your back of the envelope math?
Pocket and VPN make money, that would be like firing IRS auditors in the name of efficiency.
I agree that general purpose AI isn't really all that interesting, since I don't think it is going to drive involvement or investment. I also imagine that it doesn't really cost that much - they don't have any real products behind it, and they all seem clearly experimental.
I guess I understand your aversion to contributing to "junk projects", but if they are junk projects, there isn't likely to be a ton of investment. Harder to shift the bottom line.
I really am curious. I'm not a fan of AI, so I would agree that those seem superfluous -- but at the same time, the AI based image summarizer actually sounds cool - and good for accessibility. The translation service is VERY useful, and it is amazing that it runs locally.
You have to remember that sometimes when that shiny new CSS feature comes out, it is underspecced, with unhandled corner cases -- "just do what Chromium does" is not a standard -- or is it? Having multiple implementations of a spec prove that it is interoperable - without that, you might have a good spec, or you might have a spec that says "whatever Chrome does is what is expected". Not sure that is what we want from new CSS (or any) features.
The 2FA thing sounds like it's all on the Dropbox side if you are just entering a code you got from an authenticator app. The Google login issue may be a real issue -- did the Google login specifically work on another browser?
I wrote up what happened (more stuff on my blog).
TL;DR they use aggregated data for ads and they felt like they needed to have an explicit opt-in.