Storage is a commodity. Every AI product can keep your conversations; most now do. The question that matters is different: when you open a new tool tomorrow, does it know you — and who decided what it got to know?
That’s the line between memory and identity. Memory is what a product holds about you, inside itself, on its terms. Identity is yours: one understanding of your people, projects, decisions and ways of working that travels — readable by any AI you permit, sliced to what each one needs, revocable the moment you change your mind, exportable whole.
A slice is a grant, not a copy
Your coding tool gets your work context and nothing about your family. A writing tool gets your projects and nothing about money. The grant is enforced at read time — revoke it and the door closes; there is no copy left behind to wonder about. Every read lands in an audit log, so “what does this app know about me” is a page you open, not a support ticket.
Why this is the moat, and the bet
Models will keep changing. Apps will keep multiplying. Devices are coming. The one thing that shouldn’t reset at each boundary is you — and whoever holds that continuity holds real power, which is exactly why it must belong to the user rather than to any one AI vendor. An email address stopped being just for email; your identity won’t stay just for chat.
Everything else we publish here — the evidence gates, the benchmarks, the pseudonym ledger, the resolved dates — exists in service of this one bet: identity you own, above the models, for as long as you choose to carry it.