
Mowing the lawn. The roots untouched. That was where the last chapter left us — the most powerful defender in the ecosystem demonstrating it at maximum scale, in public, over more than a year. Now hold that image still and ask the question it begs. If the most muscular enforcement campaign ever run against this kind of abuse leaves the roots untouched, what exactly are the roots, and why is every blade we have — the rule engines and the shared banlists, the AI moderators reading intent, the platform's own server-side hammers swinging at half a million channels a day — incapable of reaching them?
This chapter answers that. It is the hinge of the book, and it does not add another defense to the pile. It steps off the treadmill entirely and asks why the treadmill exists. The whole arsenal we have surveyed — classical and AI, client-side and server-side, single-room and federated — is serious work by serious people, and a great deal of it functions exactly as advertised.
And none of it touches the thing that is actually wrong.
What the chapter covers
- The shape of the limit
- Douceur, 2002: the names are free
- EigenTrust, 2003: ban the actor, not the artifact
- Symptomatic medicine, for twenty years
- Why "detect harder" cannot escape it
- The variable nobody is touching
- The question we have refused to ask
Terms explained in this chapter
The full chapter is in the book.
Buy the full book