Yeah, I worry about that. Seems like everything a human can do to prove they're human, imitation machine learning will be able to look at those inputs and outputs and fake it.

If you can't have spaces filter out bots (already hard, probably impossible in the future), trust and validity completely go away. Spam will become unmanageable, and in many cases more subtle. Abuse will scale and could become unblockable.

I'm sure there sill be some kind of good ideas to push back, but I'd certainly say I'm pessimistic atm.