New Federated Protocols

I’ve been thinking a lot lately about a new form of software, or really an old form of software called federation, which allows for distributed protocols across various applications. Given my background at Coinbase, it’s easy to instantly associate this concept with something like a blockchain. Candidly, blockchain is one of the first mainstream examples of a federated protocol operating successfully at scale.

But there are other examples right in front of us that most people in my generation don’t even recognize as federation. HTTP is a federated protocol. SMTP, the underlying engine for all email, is another. Even text messaging relies on its own shared protocol languages.

What gets fascinating is how the rise of vibe coded applications opens the door for entirely different UIs to solve the exact same problem set. Take calendar applications, for instance. Every calendar needs a way to expose available times, schedule appointments, and capture form fields like meeting locations or map links. None of these features are unique to any single provider; Google Calendar, Calendly, and every other major player all share these same basic requirements.

If you were to expose these primitives outside of the application layer as a standard schema, you could create a protocol that lets anyone schedule on anyone else’s calendar. You could even build a system where AI agents handle the scheduling for you.

This isn’t just about the tech giants like Google. Imagine someone wanting to vibe code their own version of Google Calendar because they want a custom UI or specific form elements for a unique signup paradigm. Instead of relying on a centralized service, they could just vibe code the interface and plug into the federated schema and protocol. Their creation would be completely accessible via the traditional UI, but also wide open via API and MCP for agents, which is incredibly cool.

Ultimately, protocols form an abstraction layer on top of basic primitives. Email is the perfect historical example of this. So many massive companies have been built on top of email simply because federation kept the access open.

For the past decade and a half, the default software model has relied on centralized players building and locking down the business logic. But we might be heading toward a vast federation where individuals can choose to sign up for existing services or quickly build their own using agents. Because the underlying protocol handles the heavy lifting, all of that complex building logic is completely abstracted away from the user.

The early internet was full of this. We had a wealth of protocols and schemas, like RSS for news and blogs, which faded away as centralized companies took over. But as vibe coding gets cheaper and the unit economics become undeniable, there is a real opportunity for a huge swath of scaled applications to switch back to a federated model. It is a massive win for users because it is cheaper, highly customized, and self-sovereign. You are no longer beholden to any platform. You own it yourself, and that is a beautiful thing.