The current wave of AI agents is diminishing the value of identity as a DDOS or content-moderation signal. The formula until now included bot = bad, but unless your service wants to exclude everyone using OpenClaw and friends, that's no longer a valid heuristic.

If identity is no longer a strong signal, then the internet must move away from CAPTCHAs and logins and reputation, and focus more on the proposed content or action instead. Which might not be so bad. After all, if I read a thought-provoking, original, enriching comment on HN, do I really care if it was actually written by a dog?

We might finally be getting close to https://xkcd.com/810/.

One more half thought: what if the solution to the Sybil problem is deciding that it's not a problem? Go ahead and spin up your bot network, join the party. If we can design systems that assign zero value to uniqueness and require originality or creativity for a contribution to matter, then successful Sybil "attacks" are no longer attacks, but free work donated by the attacker.