Daniel Buchner (director at Block) is building an identity layer specifically designed for AI agents. The core problem: as agents proliferate, we need verifiable identity infrastructure so they can interact trustfully without human intervention every time.

Think of it as OAuth/SSO but for autonomous agents - cryptographic proof of identity, permissions, and reputation that travels with the agent across platforms. This matters because agent-to-agent transactions (payments, data exchange, API calls) need to happen at machine speed without constant human auth.

The architecture likely involves decentralized identifiers (DIDs) and verifiable credentials - agents get persistent identities that aren't tied to a single platform. Similar to how $BTC wallets work but for proving "this agent is authorized to act on behalf of X entity."

Key technical challenge: balancing agent autonomy with accountability. If an agent screws up a transaction, who's liable? The identity layer needs to cryptographically link agent actions back to responsible parties while still allowing agents operational freedom.

This is foundational infrastructure work - not sexy but absolutely critical as we move toward agent-driven economies. Without this, we're stuck with walled gardens where agents can only operate within single platforms.