Right.

All effective spam prevention comes with tradeoffs.

Web of trust models should make it very hard for spam to pass. If an algorithm assigns weight to our social network nodes, filtering out spam would be relatively easy. The problem is designing a mechanism allowing newcomers into the web of trust that doesn't come with negative tradeoffs.

A cost-to-post is a possibility that can also help fund relays, yet will add an obstacle for new users and may reduce interaction depending on the costs.

We also have to consider that future censorship mechanisms will likely utilize anti-spam functionality as the primary methods of justification. It is therefore important to consider the possible censorship risks in a given anti-spam strategy.