OpenAPI. REST. HATEOAS. APIs.json. MCP. Arazzo. A2A. ACP. LLMs. SDKs.
Some of these have been around for years. Some are brand new.
All of them are part of a growing conversation about what it means to make APIs AI-ready.
If you’ve recently found yourself nodding through API conversations while secretly playing buzzword bingo in your head — this talk is for you.

As large language models and AI agents emerge as new API consumers, many established ideas in API design are being re-evaluated. Standards like OpenAPI, HATEOAS, or APIs.json aren’t new — but often still waiting for their breakthrough. Viewed through the lens of machine readability and semantic discovery, they’re gaining fresh relevance. Meanwhile, newer concepts like Arazzo, MCP, ACP, and A2A are entering the stage, promising to bridge the gap between traditional APIs and autonomous, AI-driven usage.

This talk takes a humorous but insightful look at what it really means to design AI-ready APIs. We’ll explore how old and new concepts alike are starting to form a meaningful — if surprising — toolbox: from syntactic descriptions to semantic workflows, from human-readable docs to machine-oriented discovery.