I see two futures ahead: one with "free" content (data harvesting) remaining alive through remote attestation, physical key verification, phone verification etc. and one with completely paid and exclusive communities scattered around with only a few percent being able to access a meaningful amount of information. Maybe both. But things dont seem to be as bright as some AI lovers make it to be. Hopefully im just being unrealistically pessimistic and open governance prevails, somehow.

I would be thrilled about the end of free accounts. Things that don’t seem to cost that much to run can charge token amounts, and things that cost more like say, Gmail, should just cost money. Right now the existence of the shitty, ad-supported version of everything drives out anything good. Why build and innovate in any consumer software product when Google is there offering a free ad-based one that will always get 90+% of the users?

Captchas have never been reliable, the whole point was just to have a mechanism that costs more to decrypt than to produce.

I think we're still there as the cost of running the models stays high, though it's subsided at this point. And I don't if we'll ever hit a point where decrypting and encrypting costs reverse.

Captchas are used to distinguish between anonymous humans and scripts trying to impersonate a large quantity of unrelated humans. The other way of preventing someone from impersonating many fake people with a script is requiring actual verification of identity - that's a higher barrier of entry and (for now) requires country-specific solutions and so is harder to internationalize/globalize, but this seems the direction the world is moving towards.