MCP, LLMs, and the Docker Model Runner – The Open Letter, Edition #3
Hey, welcome to the 3rd edition of ‘The Open Letter’! I hope you’re all doing well and either working towards the goals you’ve set for 2025 or just going with the flow like me. I know this publication took way longer than expected, but you know—life always surprises you with different challenges.
Announcements:
So, first, let’s take a look at the cool things that happened in the past weeks and last month:
Docker announced their new offering, Docker Model Runner, which allows you to run LLMs locally. As of writing this newsletter, it’s supported on Macs with Apple Silicon chips, with Windows support coming soon. The Docker team is also planning to open source it.
There’s a lot of conversation around the Model Context Protocol (MCP), and it’s shaping up to be the next big thing in the AI space. In simple terms, with an MCP server, we can give LLMs more context capabilities to get better and more real-time answers.
Git turned 20 (no longer a teenager), and it’s one of those tools you can’t live without.
In the Large Language Model war, Meta released Llama 4 with significant improvements and added support for multimodal capabilities. OpenAI also released GPT-4.1, which excels in coding, follows instructions well, and handles long contexts of up to 1 million tokens. On the other hand, Google launched Gemini 2.5, claiming it is their most intelligent AI model.
Resources:
I recently published a blog on running LLMs locally with Docker. In less than 72 hours, it crossed 10k views and was trending at the top under the Docker tag. A lot of enthusiasm there.
Git recently turned 20. On this occasion, GitHub had a conversation with creator Linus Torvalds
Conferences and CFPs:
Thoughts:
That brings us to the end of this newsletter. Again, thank you for reading the third edition, subscribing to it and making it to the end. If you want to connect with me or have any feedback to share, I would love to hear that. Here is the link. — Pradumna Saraf
Amazing update pradumna, Thanks for these updates!
awesome pradumna