🤖 AI Summary
Overview
This episode explores the emerging Model Context Protocol (MCP), a system designed to enable AI to interact with the internet as humans do. Alex Rattray, CEO of Stainless, shares insights on building MCP servers, the challenges of scaling them, and their potential to revolutionize AI-driven automation. The conversation also delves into the future of AI, security concerns, and practical applications of MCP in business operations.
Notable Quotes
- The longer you wear your shoes, the more worn out they get, but the longer you just wear your feet, the tougher they get.
– Alex Rattray, on his unconventional barefoot running philosophy.
- The future of AI is cyborgs—part neural net, part traditional code.
– Alex Rattray, envisioning the hybrid nature of AI systems.
- We haven’t figured out how to expose an API ergonomically to an LLM in the same way we’ve cracked it for Python developers. That’s a new research problem.
– Alex Rattray, on the challenges of designing MCP servers.
🧠 APIs and MCP: The Connectors of the New Internet
- Alex Rattray explains APIs as the dendrites of the internet,
enabling software to communicate seamlessly.
- MCP (Model Context Protocol) is introduced as a way for AI to interact with websites and services, akin to how humans use user interfaces.
- MCP servers aim to expose tools to AI, allowing it to perform tasks like sending emails or processing refunds autonomously.
⚙️ Why MCP Servers Are Hard to Get Right
- MCP servers face scalability issues due to the limited context windows of language models (LLMs).
- Exposing too many tools overwhelms the model, while too few tools restrict functionality.
- Security concerns arise when AI tools have broad access to APIs, risking unintended actions like mass refunds or data breaches.
- Alex Rattray emphasizes the need for precise tool descriptions and minimal input/output data to optimize performance.
📊 Using MCP for Business Operations
- Stainless leverages MCP servers for tasks like querying databases, analyzing customer data, and cross-referencing information across tools like HubSpot and Notion.
- Alex Rattray uses Claude Code to build a company brain,
storing SQL queries, customer quotes, and analytics in a Git repository for future use.
- This approach reduces repetitive work and enables faster decision-making by centralizing knowledge.
🔒 Security and the Future of MCP
- Security for MCP servers should be enforced at the API layer using OAuth with granular permissions and scopes.
- Stainless is exploring ways to make this process easier for developers while maintaining robust security.
- The vision for MCP includes AI writing and executing code directly, reducing reliance on predefined tools and enabling greater flexibility.
🚀 The Future of AI: Cyborgs and Code Execution
- Alex Rattray predicts AI systems will combine neural networks with traditional code execution environments, creating cyborgs
that seamlessly integrate both.
- Code execution tools will allow AI to write and run scripts dynamically, interacting with APIs more efficiently than current MCP setups.
- This shift could simplify tool creation, with prompt engineering becoming the primary focus for developers.
AI-generated content may not be accurate or complete and should not be relied upon as a sole source of truth.
📋 Episode Description
If your MCP server has dozens of tools, it’s probably built wrong.You need tools that are specific and clear for each use case—but you also can’t have too many. This creates an almost impossible tradeoff that most companies don’t know how to solve.
That’s why we interviewed Alex Rattray, the founder and CEO of Stainless. Stainless builds APIs, SDKs, and MCP servers for companies like OpenAI and Anthropic. Alex has spent years mastering how to make software talk to software, and he came on the show to share what he knows. We get into MCP and the future of the AI-native internet.
If you found this episode interesting, please like, subscribe, comment, and share.
Want even more?
Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.
To hear more from Dan Shipper:
- Subscribe to Every: https://every.to/subscribe
- Follow him on X: https://twitter.com/danshipper
Ready to build a site that looks hand-coded—without hiring a developer? Launch your site for free at Framer.com, and use code DAN to get your first month of Pro on the house.
Timestamps:
00:00:00 - Start
00:01:14 - Introduction
00:02:54 - Why Alex likes running barefoot
00:05:09 - APIs and MCP, the connectors of the new internet
00:10:53 - Why MCP servers are hard to get right
00:20:07 - Design principles for reliable MCP servers
00:23:50 - Scaling MCP servers for large APIs
00:25:14 - Using MCP for business ops at Stainless
00:28:12 - Building a company brain with Claude Code
00:33:59 - Where MCP goes from here
00:41:10 - Alex’s take on the security model for MCP
Links to resources mentioned in the episode:
- Alex Rattray: Alex Rattray (@RattrayAlex), Alex Rattray
- Stainless: https://www.stainless.com/