What should an ideal digital identity system look like?

In this episode of Coordinate, Soubhik sits down with dcbbuilder, Research Engineer at World Foundation, to unpack the architecture of digital identity from first principles.

We start with the core mental model: credentials, issuers, roots of trust, attestations, and relying parties. From there, we dive into the properties of a strong identity system, privacy, unlinkability, recovery, revocation, self-hosting, and programmable attestations.

In the second half, we go deep into World ID’s design: how the Orb works, how privacy is preserved, why client-side proving matters, how iris uniqueness checks use MPC, and how zero-knowledge proofs enable proof of personhood without revealing identity.

We also explore one of the biggest open questions ahead: identity for AI agents, what it means, why it matters, and how human-linked agents may reshape trust online.

A great episode for anyone interested in digital identity, zero-knowledge proofs, privacy tech, biometrics, MPC, and the future of the internet.

Timestamps

00:00 Introduction

00:13 Why digital identity matters

01:24 The mental model: credentials, roots of trust, and relying parties

04:28 Can roots of trust be decentralized?

06:56 What an ideal digital identity system should look like

07:28 Privacy, unlinkability, and why they matter

13:21 Recovery, revocation, and self-hosted identity

18:38 Programmable attestations, ZK, MPC, and TEEs

36:30 How World ID works: Orbs, privacy, iris codes, and MPC

47:54 World Chain, mini apps, and identity for agents

Reference links mentioned in the Podcast:

- Future of Digital Identity 1: https://dcbuilder.dev/blog/the-future-of-digital-identity

- Future of Digital Identity 2: https://dcbuilder.dev/blog/a-tidal-change-the-future-of-digital-identity-ii

- AMPC Privacy/ World ID: https://world.org/blog/engineering/introducing-ampc-another-leap-privacy-performance-world-id

- Learn about World Orb: https://world.org/blog/world/orb-faqs