Open, anthropic, and block combine to make AI Agents play nice

Open itIt is anthropic, and the block has released a new open source organization – the agentic ai Foundation – to promote artificial intelligence tax standards.

The three companies have also transferred ownership of some of the most widely used agentic technologies to the base. These include the anthropic model Openai’s Agents.MD, which allows programs and websites to specify the rules of the code agents; And goose, a framework for building agents developed by Block. These technologies were already free to use, but with a new foundation it will be possible for others to contribute to their development.

“MCP is used by many companies, but there are others [who don’t use it].

The Agentic AI Foundation is created under the Linux Foundation, responsible for the development of the widely used Linux operating system LINUX LINEX and other projects. The foundation provides legal and technical support for open source infrastructure development. Other companies that have signed up to Aaif, beyond the three founding members, include Google, Microsoft, AWS, Bloomberg, and CloudFlare.

The new foundation shows a nascent shift from risk-based AI systems to the greater use of systems that take action on behalf of users. This type of agentic AI promises a potentially profitable paradigm where AI agents use the web and communicate with each other to serve all kinds of requests. Consumers, for example, use AI assistants to purchase and book items, while businesses use AI Agents to manage transactions and customer interactions.

Srinivas Narayanan, chief technology officer of B2B applications at Opelai, sees a time when a large number of AI Agents are constantly communicating within the business. An AI industry that works across the same open standards should help ensure that those interactions happen seamlessly. “Open source will play a huge role in how AI is shaped and adopted in the real world,” Narayanan said.

The question of openness seems important to Ayi right now. US companies mostly make money by providing access to closed dynamic models through application programming services, or APIs. Meta previously pre-released its flagship model, the llama, so that anyone could pick it up and run it, although the company recently signed off on a closed-door replacement. Many Chinese AI companies, including Deepseek, Alibaba, Moonshot AI, and Z.AI, offer open source models that are popular with developers, first timers, and AI researchers. Some worry that this image could give Chinese firms a bigger advantage in the long run.