There are 3 problems with this statement.
1) There is a lot of free content, easily found, and content gets more rapidly out of date nowadays. Paid content must deal with this (newsworthy, in depth competition).
2) Your view is US centric; the vast majority (read: not all) of internet users from other countries generally have less money to spend.
3) The cost of loss of privacy is indirect, therefore hidden, and US companies generally get away with it. This is why data mining is currently profitable.
In order to solve #3 shit must first hit the fan, to increase awareness. #1 is solved by collectively putting the high quality content behind paywall. #2 can in theory be solved by globalisation in long-term. Short-term, dynamic price makes sense, but it makes #1 & #3 worse.
Let me quote the rest of your statement:
> [...] They want stuff for free. And if site A starts charging, they will jump ship to site B, until eventually it's a race to the bottom.
Agreed, except for example for (local) newspapers who also provide online access as part of their sub. Global newspapers have more competition. Especially given the widespread English language.
> Ads seems to be the only model that works (thus far).
For Google, yes. For privacy, no. In apps on an unrooted Android device, yes. On recent iOS and desktops, no (the latter 2 due to ad blockers).
Also, streaming media such as Spotify proves a subscription in a new market (not newspapers) can work. Netflix proves it for video.
> And recently Mozilla's Brendan Eich is doing Brave, with bitcoin micropayments. I'd love to see it gain traction, but I don't see it going mainstream.
Not yet a final, stable non developer version released AFAIK. Also, Eich is no longer part of Mozilla.