Training Data
Training Data

Anthropic's Katelyn Lesse & Angela Jiang: Building an Ecosystem, not a Walled Garden

July 14, 2026

AI Summary

5 min read

Building an Ecosystem, not a Walled Garden

Anthropic's platform team operates with a dual mandate: serve external developers building on Claude's APIs while simultaneously powering Anthropic's own internal products. Katelyn Lesse and Angela Jiang, who lead this effort, describe their North Star as giving "any builder the tools to work with Claude to build whatever they want to build." This means providing primitives, higher-order abstractions, and open standards rather than locking developers into a single product. The team deliberately keeps internal and external offerings on the same infrastructure, reasoning that form factors evolve too quickly to bet on any single one. "Two years ago we were all like everything's chat, and now everyone's like forget chat and just agents," Angela notes. The only way to keep up is to build a robust platform that lets the market discover form factors rather than dictating them.

The Three-Layer Cake of Abstractions

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

What you'll learn

  • 1 Timestamped Navigation Outline
  • 2 (00:00) **Three-Layer Abstraction Cake** - Angela introduces the knowledge, execution, and coordination layers that structure Anthropic's platform roadmap
  • 3 (01:04) **Platform's Two North Stars** - Caitlin and Angela define their team's dual mission: internal speed for Anthropic's own product teams and external empowerment for all builders
  • 4 (03:54) **Internal vs. External Platform Philosophy** - The team intentionally keeps internal and external primitives equal rather than bifurcating them
  • 5 (06:14) **From Stateless API to Higher-Order Abstractions** - Caitlin traces the evolution from a simple messages API to solving customers' repeated infrastructure problems
  • 6 (08:29) **Who Uses Primitives vs. Managed Offerings** - Usage splits cleanly between AI-native startups and everyone else
  • 7 (09:13) **Three-Layer Cake Deep Dive** - Angela walks through the concrete primitives at each layer: knowledge, execution, and coordination

+ Full timestamped outline available in the app

Show Notes

Katelyn Lesse and Angela Jiang lead the team building Anthropic's developer platform - the layer that both outside builders and Anthropic's own products run on top of. Angela frames the platform as a three-layer stack: knowledge, execution, and coordination. She argues the real leverage is what’s at the top: "strategies," or meta-harnesses that give each token a different job, from advising to executing to reflecting to memory. On the question of open ecosystem vs. walled garden, they say they aren't precious about owning the stack. Katelyn points to Anthropic's self-hosted sandboxes with partners like Modal, Vercel, and Cloudflare. Whether the work runs on Anthropic's infrastructure or someone else's, what really matters to them is that the architecture is sound. The deeper bet is standards: they hand skills and MCP to the whole industry, build connectors on the MCP spec, and help agents (Claude and non-Claude) work together. The one place they stay closed is model routing: they argue harnesses should be tuned to a model family, so they're designing for Claude rather than routing across models. Angela's frame for the ecosystem bet is electricity: transformative only because everyone could plug in, and no company wired it alone.
Hosted by Sonya Huang and Lauren Reeder, Sequoia Capital


00:00 Introduction

01:49 Two North Stars

02:27 External Builders And Primitives

03:54 What To Externalize

06:00 From Messages To Agents

08:19 Managed Agents Adoption

09:07 Three Layer Cake

10:22 Execution Harnesses Explained

11:09 Coordination Strategies Roadmap

12:13 Ecosystem Standards And Safety

15:39 Open Ecosystem Not Walled

17:12 Vertical Products And Form Factors

22:26 Claude Tag Under The Hood

26:04 Harness Best Practices

38:13 Token Costs And Whats Next



Training Data

More from this podcast

Training Data →