OpenAI, Anthropic, and Block join new Linux Foundation

pull down to refresh

OpenAI, Anthropic, and Block join new Linux Foundation | TechCrunch techcrunch.com/2025/12/09/openai-anthropic-and-block-join-new-linux-foundation-effort-to-standardize-the-ai-agent-era/

328 sats \ 5 comments \ @Scoresby 21h AI

The group, dubbed the Agentic AI Foundation (AAIF), will act as a neutral home for open source projects related to AI agents. Anchoring the AAIF at launch are donations from Anthropic, Block, and OpenAI.

Anthropic is donating its MCP (Model Context Protocol), a standard way to connect models and agents to tools and data; Block is contributing Goose, its open source agent framework; and OpenAI is bringing AGENTS.md to the table, its simple instruction file developers can add to a repository to tell AI coding tools how to behave.

Other members in the AAIF include AWS, Bloomberg, Cloudflare, and Google, signaling an industry-level push for shared guardrails so that AI agents can be trustworthy at scale.

“By bringing these projects together under the AAIF, we are now able to coordinate interoperability, safety patterns, and best practices specifically for AI agents,” Zemlin said.

I am but a young girl, and unschooled in the ways of war, but I don't really understand why this is necessary. Why are open source licenses not enough? Sounds like mumbo jumbo to me, but perhaps it is just my ignorance showing.

view all related items

177 sats \ 2 replies \ @optimism 21h

It's capture: you formalize open source frameworks then you have to worry less about disruption.

100 sats \ 1 reply \ @BlokchainB 17h

Disruption from a competitor?

287 sats \ 0 replies \ @optimism 12h

From a FOSS developer? A Chinese competitor? A government trying to pick winners?

I don't think that this is the rationale for Block or the Linux Foundation; I roughly believe their narrative. But for the others? There's a little hint at the bottom of the article:

There’s also a more subtle consequence: Even with open governance, one company’s implementation could become the default simply because it ships fastest or gains the most usage.

So why is Anthropic donating MCP? Funnily, I still have my Claude Code setup which I play with sometimes and there isn't a single MCP server since 4.5 was released. It's not needed anymore: someone else's middleware will perform poorer against my prompts than the tools I let Claude develop for my prompting style. Someone bound to MCP is someone bound to lose in a race.

AGENTS.md from OpenAI is even more questionable; Anthropic has argued against using it that way recently because it often poisons context and I must agree with them on that. They've also introduced skills (which is counter to the single agent textfile) for their front-end software. Cursor/Cline/Roo have also had an improved concept for this since forever.

Also note these are poor standards and I've always suspected that they've been vibed and not designed by an experienced protocol/standards developer.

And then, the example straight after makes it so bad, it's funny:

Zemlin says that’s not necessarily a bad thing, though. He points to open source history — like Kubernetes “winning” the container race — as evidence that “dominance emerges from merit and not vendor control.”

Anyone that in 2024/2025 tried to provision a kubernetes cluster from scratch on sovereign compute knows that this is fully optimized for big centralized cloud providers now. It's a great example of something that's been captured.

0 sats \ 1 reply \ @anon 20h

I'd like to tell you about our lord and savior, Haiku OS>

0 sats \ 0 replies \ @SwapMarket 19h

Wow, the UI reminded me OS/2 from the 90s, before we all switched to Windows. I did not hear about BeOS or Haiku. Does it have any games? BTW, none of the linked youtube videos work.