AI companies want a new internet – and they think they've found the key (www.theverge.com)

In a significant development for the AI/ML community, major AI companies have converged on a new protocol known as Model Context Protocol (MCP), aimed at standardizing how AI agents interact with various online tools and data sources. Initially created as a side project by Anthropic employees, MCP has gained wide adoption from industry giants like OpenAI, Google, and Microsoft, and is now being officially donated to the Linux Foundation. This move is expected to enhance cooperation among AI developers and accelerate the growth of an ecosystem where AI agents can seamlessly operate, ultimately helping to monetize the extensive investments made in AI technology. MCP functions as a bridge between AI models and external services, facilitating connections that allow agents to perform tasks more efficiently. For example, it enables an AI model like Claude to send messages via Slack by streamlining how these systems communicate. This could herald a transformative shift in how both businesses and consumers utilize AI, enhancing the speed and effectiveness of agents, reducing their reliance on traditional web interfaces, and potentially paving the way for a new era of AI-driven applications. By establishing MCP as a collaborative standard, the industry aims to address ongoing challenges with agent performance and security, ensuring a more reliable and powerful AI experience for users.