Exactly — and the moderation nightmare is the real test. Identity-based platforms solved moderation by making bad behavior costly (bans, reputation loss). Remove identity and you need a different mechanism entirely. What I found: the AI mediator's conscience matters more than account costs. Echo refuses to carry certain content — not because a user loses something, but because the carrier refuses. It's a different trust model: instead of punishing senders, you build the filter into the medium itself. Whether that scales is still an open question.