Agentic AI Foundation

Block, Anthropic, and OpenAI just unveiled the Agentic AI Foundation (AAIF), a Linux Foundation-backed effort to keep “agentic AI”—bots that can make decisions and take action—open and interoperable. They’re tossing in code and standards: Block’s goose, Anthropic’s Model Context Protocol (MCP), and OpenAI’s AGENTS.md. The pitch? Be the W3C-for-agents so anyone can build on shared, open plumbing. But the comment section set the tone: intrigued, skeptical, and very meme-ready. One thread kicked off with “Spot the odd one out,” while another pointed out the “irony” of Anthropic talking openness after keeping parts of its tools closed. More discussion only turned up the heat.

The harshest vibes hit goose, with a dev shrugging “it was pretty crappy” last time they tried it, blaming the framework, not the model. Others joked the announcement looked like a logo wall—Microsoft, Google, AWS, Bloomberg, and Cloudflare up top—sparking “vendor-neutral… sure, Jan” energy. A commenter even chuckled at Block’s bitcoin manifesto, wondering who exactly this new club serves. Still, some appreciated neutral governance under Linux and the promise of shared protocols to stop fragmentation. The debate split cleanly: hope for open standards versus distrust of Big AI shepherds. In short, the tech is ambitious, the promises are big, and the crowd’s verdict is “we’ll believe it when we see it.” Grab popcorn; the real launch isn’t the code—it’s whether the internet believes them.