We are at a crossroads. The next age of humanity will be shaped by AI, identity systems, and digitized trust — the only question is who shapes them, and whose interests get encoded into the substrate.
Right now, the direction of travel is bad. Trust in institutions is in decline. Inequality is being hard-coded into the platforms we depend on. AI capability is concentrating in the hands of a few entities with the compute to train and serve it. And on the open internet, we are losing the ability to tell who — or what — we are talking to. Human? AI? Whose AI?

These aren't separate problems. They share a root cause: we've outsourced the production of trust to centralized intermediaries, and those intermediaries are now also the gatekeepers of computation, capital, and identity. The same actor verifies you, ranks you, and runs the model that talks to you. That is not a stable arrangement, and it is not one we should want to stabilize.

The alternative isn't another platform, and it isn't anonymity. It's a decentralized trust network — a base layer where humanity can be proven, identity is held locally, and trust propagates through a graph that humans themselves shape.

Concretely: Alice gets a humanity attestation from her local trust circle — people who know her, vouch for her, whose own attestations form a web she carries into the network. With it, she delegates a personal AI agent that carries her identity outward. It's her agent. It's bound to a human. And when it speaks to yours, both sides can verify there's a person on the other end of the chain.

This is where topology becomes the design surface. A trust network is a graph, and the shape of that graph determines what the system can do — how attestations compose, how sybil resistance emerges from social structure rather than central authority, how locality and delegation work. Letting topology emerge by accident gives you the social graphs we already have, optimized for capture. Designing it intentionally gives you a substrate other networks can safely build on.

A mesh of AI agents anchored to people becomes a real counterweight to centralized model providers. Funding networks can route capital through human-attested graphs instead of platform algorithms. Reputation, coordination, governance — all of it can sit on a base layer that wasn't designed to extract from its users.

We're here because the people in this room are the ones who can build this substrate — or watch something worse get built by default.

Trust is infrastructure now. Decide what shape you want it to have, before someone else decides for you.