OpenAI recently incorporated support for the Model Context Protocol (MCP) into its Agents SDK. MCP is positioned as a standardized open protocol that facilitates the connection of applications to large language models (LLMs) by defining specific types of servers based on their transport mechanisms. The main intention behind MCP is to provide context to LLMs in a more uniform manner. While some users see potential in MCP for creating streamlined connections akin to USB-C for AI applications, others are skeptical and argue that existing solutions like OpenAPI already handle these needs effectively. Feedback from the developer community reflects concerns about MCP's flexibility and documentation, with some expressing frustration that it complicates an already established process rather than enhancing it. Additionally, there are hints of features like OAuth 2.1 and JSON-RPC Batching being introduced, which may contribute to the protocol's utility. Overall, despite some endorsement from OpenAI, the community remains divided on the actual benefits and practical implementation of MCP in connecting LLMs to external tools, with many stressing that it requires further exploration and development.