Over the last year, MCP servers have become to standard for connecting our AI agents to our tools and data. You've got MCP servers for Google Drive, Slack, Notion, Playwright, Archon, Postgres, and the list goes on and on. If there's a tool, there's an MCP for it. But it's not like you can use all of these MCP servers together. If you tried, you would completely overwhelm your agent. And therein lies the problem with MCP - each one of these servers eats up a TON of context for your LLM (context rot) and also is rather inflexible. In this video I break down the problem super clearly and super fast. Then I go over the solution (where we are heading) with an example using my open source tool Archon! ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - If you want to master AI coding assistants and learn how to build systems for reliable and repeatable results, check out the new Agentic Coding Course in Dynamous: - Anthropic's article on the problem with MCP (and solutions): - Claude Skills: - Archon Skill: - Archon: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 00:00 - MCP Popularity has Exploded, but is it Justified? 00:58 - The Big Problem with MCP 01:26 - Example for How Bad MCP "Context Rot" Can Get 02:25 - MCP Problem Demonstrated in Claude Code 02:59 - The Solution to the MCP Problem 04:51 - How Agent Code Execution Works 05:46 - Claude Agent Skills - The Solution Made Easy 08:08 - Is This the End of MCP? 09:43 - Final Thoughts ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Join me as I push the limits of what is possible with AI. I'll be uploading videos weekly - at least every Wednesday at 7:00 PM CDT!