Re: Discussing the Decision to Not Make FastMCP the MCP Python standard

[gitanon112]

Hey @jlowin,

Just wanted to say thank you very much for your great work, FastMCP is by far the superior approach to MCP, and it is really disappointing that the official MCP python-sdk has dropped the ball in so many ways, when the best decision IMO would be to simply make FastMCP the standard. I actually made an issue discussing this in python-sdk repo.

Honestly, I am just curious about the why? IMO it seems like the pydantic/pydantic-ai team has got executive control over the official MCP python, and they are just not great at it. Pydantic-ai is already obsolete IMO like langchain, heavily over-bloated/complex, and they are simply not really able to grasp the future potential of MCP coherently.

Sorry, I do not want to stir drama at all, and they have done some great work, but it really feels like whoever is in charge of it is only limiting/slowing down AI progress. They are stuck in repetitive, fruitless rumination cycles, when they should just be shipping(recently they have started again). Just curious about your thoughts, sorry if this is the wrong place for this, just frustrated.