

WTF is MCP? w/ David Cramer
Apr 24, 2025
Dive into the innovative world of microservices for language models with insights from a special guest. Discover how AI agents are shaping the future of development and why job security isn't a concern with emerging tools. The hosts share their whiskey tasting experiences while engaging in playful debates over frameworks like Rails and Laravel. Expect captivating anecdotes about life, tech, and even the challenges of driving in hilly cities, sprinkled with humor and expert insights.
AI Snips
Chapters
Books
Transcript
Episode notes
MCP Explained with Sentry Example
- David Cramer describes MCP as microservices designed specifically for language models by hosting functions remotely.
- He illustrates this with Sentry's patch generation, combining various tool calls into reusable services for AI agents.
MCP's Efficiency in API Design
- MCP allows efficient, composable AI tooling by exposing APIs tailored for language models instead of large JSON blobs.
- This specialization overcomes traditional internet API inefficiencies and supports better interoperability in B2B.
AI Aids, Doesn't Replace Developers
- AI-generated code tools currently speed up development by 2-3x but do not replace developers.
- True productivity gains come from people leveraging AI effectively, not AI alone replacing jobs.