scripod.com

The Creators of Model Context Protocol

Shownote

Today’s guests, David Soria Parra and Justin Spahr-Summers, are the creators of Anthropic’s Model Context Protocol (MCP). When we first wrote Why MCP Won, we had no idea how quickly it was about to win. In the past 4 weeks, OpenAI and now Google have now announced the MCP support, effectively confirming our prediction that MCP was the presumptive winner of the agent standard wars. MCP has now overtaken OpenAPI, the incumbent option and most direct alternative, in GitHub stars (3 months ahead of conservative trendline): For protocol and history nerds, we also asked David and Justin to tell the origin story of MCP, which we leave to the reader to enjoy (you can also skim the transcripts, or, the changelogs of a certain favored IDE). It’s incredible the impact that individual engineers solving their own problems can have on an entire industry. Timestamps 00:00 Introduction and Guest Welcome 00:37 What is MCP? 02:00 The Origin Story of MCP 05:18 Development Challenges and Solutions 08:06 Technical Details and Inspirations 29:45 MCP vs Open API 32:48 Building MCP Servers 40:39 Exploring Model Independence in LLMs 41:36 Building Richer Systems with MCP 43:13 Understanding Agents in MCP 45:45 Nesting and Tool Confusion in MCP 49:11 Client Control and Tool Invocation 52:08 Authorization and Trust in MCP Servers 01:01:34 Future Roadmap and Stateless Servers 01:10:07 Open Source Governance and Community Involvement 01:18:12 Wishlist and Closing Remarks

Highlights

This podcast delves into the Model Context Protocol (MCP), a groundbreaking standard for AI applications developed by Anthropic. Created by David Soria Parra and Justin Spahr-Summers, MCP has rapidly gained traction, surpassing competitors like OpenAPI in popularity and functionality. The discussion explores MCP's origins, its technical underpinnings, and how it enhances AI capabilities through innovative integrations.
00:00
Everyone's now on the MCP bandwagon
01:46
MCP is designed like a USB-C port for AI systems
03:08
MCP started with internal developer tooling at Anthropic in July 2024.
06:24
Project started after David's pitch in July
24:32
LLMs should process raw data to maximize flexibility
30:26
OpenAPI is somewhat stateful but more statefulness will be popular for AI applications.
40:59
The concept of combining different MCPs into a super MCP is intriguing
41:13
Model independence allows flexibility in LLM summarization
42:06
Building graphs of MCP servers that can interact richly.
48:51
Core MCP principle involves client, application, and user involvement.
49:41
Tools in MCP should be model-controlled.
1:00:19
MCP allows building servers to optimize model answers after multiple tries.
1:09:40
Separate servers suggested for different authorizations
1:13:03
Governance in open source should balance openness with agility.
1:17:45
MCP client or server integrated with Godot Engine for AI game playtesting

Chapters

Model Context Protocol - David + Justin
00:00
What is MCP?
01:46
The Origin Story of MCP
03:08
Development Challenges and Solutions
06:23
Technical Details and Inspirations
09:05
MCP vs Open API
30:26
Building MCP Servers
33:26
Exploring Model Independence in LLMs
41:10
Building Richer Systems with MCP
42:06
Understanding Agents in MCP
43:41
Client Control and Tool Invocation
49:33
Authorization and Trust in MCP Servers
52:27
Future Roadmap and Stateless Servers
1:01:42
Open Source Governance and Community Involvement
1:10:02
Wishlist and Closing Remarks
1:17:45

Transcript

Alessio: Welcome back. MCP, MCP, MCP. After the NYC Summit, we wrote a popular piece explaining why we think. The Model Context Protocol from Anthropic seems to have won the Agent Open Standard Wars of 2023 to 2025. It seems everyone is now jumping on the ...