I’m missing something here—why did Docker decide to include both MCP and AI/ML? Doesn’t that go against the Linux philosophy, where each tool is supposed to do one thing and do it well?
I’m not sure I understand the question, so can you elaborate?
I guess you mean this: https://www.docker.com/products/ai-ml-development/
and this: MCP Catalog and Toolkit | Docker Docs
and here is the Model Runner documentation: Docker Model Runner | Docker Docs
Docker Inc is focusing on tools for container-based development now and people already using containers for AI and AI for development in general. Many companies add AI to their products like IDE developers and Docker Inc is doing the same.
They started with the AI integrated into the documentation so you could ask about Docker if you could not find something in the documentation alone. Then they released the currently beta Ask Gordon which could help directly from Docker Desktop and knew more than the AI i the docs. It could even interact with Docker and Kubernetes and new features were added to it.
And now you can also use MCP servers to extend Gordon’s capabilities. You can also find mcp images on Docker Hub.. Or download models as Docker images from Docker Hub.
Regarding the Linux philosophy, Docker Desktop is not a Linux app. And Docker can run Windows containers too on Windows. The main goal is a tool or tools that can help during development and is not pretty hard to install. That is actually one of the goals of containers too. Providing a software as a package that can run easily at least in some kind of demo mode until you learn to parameterize it but you don’t need to learn to install th correct dependencies. It will not always be easy of course. It depends on the image maintainer as well and what the software is.
So while Docker CE is best for running Linux containers it is based on the Moby project which allows running containers on Windows. Then Docker Desktop allows you to run Linux containers on Mac and Windows (in a virtual machine) in addition to Linux. You could install a VM manually, but it would be harder to properly configure it for development And now when AI became so popular, Docker wants to support AI development and using AI for helping with development. But these are still related to isolation and containers.