Sam Altman's World just shipped the thing everyone's been dancing around: proof that a real human authorized your AI agent to do what it's doing.
The Signal
World launched AgentKit, a toolkit that lets AI agents prove they're operating under human authority. This isn't theoretical. It's built on x402, the protocol Coinbase and Cloudflare released that gives agents a way to authenticate actions across the internet. The timing matters. We're at the point where agents are starting to handle real transactions, book travel, make purchases, manage portfolios. The question isn't whether agents will do this work, it's how we'll know who's responsible when they do.
World brings its iris-scanning Orb tech to the table for identity verification. That's the controversial part, but it's also the hard part everyone else has been avoiding. Digital signatures are great until you need to know there's actually a person behind the wallet. Coinbase adds the payments rail and protocol layer. Cloudflare adds the infrastructure that makes this work at scale across the web.
The beta release means developers can start building agents that carry verified human credentials. Not just "this wallet signed this transaction" but "this verified person authorized this agent to act on their behalf." That's a different architecture. It means agents operating with delegated authority instead of synthetic autonomy. It means when your agent screws up or commits fraud, there's a person who can be held accountable.
The Implication
Watch how quickly this becomes table stakes for agent platforms. Any service letting agents spend money or sign contracts will need human verification or face liability nightmares. If you're building agent infrastructure, the question is whether you integrate this now or build your own version later. The harder question: how long before we're all carrying World credentials just to let our agents work.
Source: Unchained