The promise of open source has always been that it’s easier for both users and vendors to adopt than proprietary options.
The latest case in point to that truism is the emergence of the Model Context Protocol (MCP). This is an open source technology developed by proprietary AI vendor Anthropic. The problem that Anthropic had was one that all AI vendors had – how to connect a model in a standardized way with external resources.
Anthropic created MCP, releasing it under the permissive MIT open source source license and it got a bunch of traction. Then it got even more. Instead of every vendor creating their own framework for interactions, which are essential for modern agentic AI interactions, vendors of all sizes have embraced MCP.
That includes even OpenAI, which for a while was going its own direction, but realized — like everyone else — that this is a layer where standardization is essential.
So now, rather than every vendor having its own specific protocol for agentic AI function interaction, there is widely adopted technology with MCP. Without open source licensing, it wouldn’t have happened quickly, easily or at all. Vendors in the AI space are fierce competitors, but here’s a nice fix for a common need and it works for everyone — without creating an unfair or differentiated advantage for anyone. Users win, vendors win.
Thanks open source!
