Essays by Ry Walker
Essays on AI, startups, and software engineering by Ry Walker, founder of Tembo and Astronomer.
On Agents
Agents Are Software
Demo to Deployment
- The Demo Is Not the Deployment
- The Operationalization Gap: Where AI Demos Go to Die
- The Mirror Problem
- Start With the Mirror, Not the Model
- Vibe Code Has No Production Strategy
- Agents in Production: GTM Mesh and the Death of the ERP
- The Pilot Is the Product
- Internal AI Tools Have a Twelve-Month Shelf Life
- Start With the Pain, Not the Platform
Build vs Buy
- Agent Build Versus Buy: Why Engineers Keep Building It Themselves
- The Agent Buyer Map: Who Builds, Who Buys
- The Agent Stack Build-vs-Buy Map
- What the Build-vs-Buy Data Actually Shows
- Homegrown Platforms Decay
- The Real Product Is What Replaces Homegrown
- Customization Is the Moat, Not Model Quality
- The Gap Is Infrastructure, Not Intelligence
- The Mega-Agent Fantasy Is Already Falling Apart
- The Procurement Trap
Harness & Factory
Context & Memory
- Context Is the Moat — Don't Give It Away
- Tech Context Is Tractable. Org Context Is Not
- Context Engineering Is the Hard Problem
- Three Layers of Agent Context, and Most Agents Have Zero
- The Organizational Context Gap
- Organizational Context Is the Hardest Problem Nobody Has Solved
- Two Maintenance Curves: Infrastructure Decreases, Context Never Stops
- Agent Memory Is Unsolved. Workflow-Scoped Learning Is Not
- Agent Memory Is the Defensible Layer
- Each Person Needs Their Own Agent Instance
- The Codebase Is the Territory. The Agent Needs a Map
- The Submodule Problem Is the Whole Problem in Miniature
- Why Sandboxes Beat Vector RAG for Code Generation
Mesh & Architecture
- The Atomic Agent Mesh: Architecture, Build-vs-Buy, and the Review Layer
- The Mesh of Specialists Pattern
- The Mesh, Not the Monolith
- The Convergent Agent Stack
- The Orchestration Bet
- Event-Driven Agents Change What Is Possible
- Triggered Workflows Generate Most of the Volume
- Background Agents Are the Underbuilt Layer
- Most APIs Are Not Ready for Agents
- The Agent Infra Maturity Gradient
Workflows & Roles
Review & Trust
- Review Is Not a Screen. It Is a Primitive
- Human Review Is Not a Limitation
- Non-Determinism Demands Human Correction Loops
- Inspectable Logic, Not Black Box Magic
- The Algorithm Should Be Inspectable
- Controllability Is Not Optional. Enterprise Teams Do Not Want Magic
- Code Review Becomes the Bottleneck
- The Agent Made a New Type Instead of Finding the Real One
- Taste Does Not Scale With Token Throughput
Coordination & Governance
Defense & Security
AI Coding Landscape
GTM & Platforms
- The Forward-Deployed Model Is the Only One That Actually Works
- From Agent to Platform: Why the GTM Is Services-Shaped
- GTM Mesh: Closing the Gap Between Ad Click and Revenue
- The Design Partner Discipline
- Follow No Goose Absolutely
- Catch the Buyers Before They Become Builders
- The Arena of Arenas: Why There Is No Winner-Take-All in Agents
- Three Audiences, Three Agent Strategies
- Think Small to Win Big
- Flexibility Is Not Optional for Agent Platforms
- The Zero-Stickiness Problem at the Tool Layer
Markets & Strategy
- The Long Game in Agent Companies
- Crowded Starting Line, Empty Finish Line
- The Token Reckoning Is Coming
- Provocatypes Over Roadmaps
- Pluck a Feather From Each Goose
- The ERP Is Dead. The Agent Is Your Operating System Now
- The Human Is No Longer the Integration Layer
- The Unit of AI Consumption Is the Organization
On Startups
Founder Notes
Decision-Making & Intuition
Tembo & the AI Era
- The Tembo Manifesto
- Rise of the Agents: An AI Coding Ecosystem Map
- Claude Code Just Got a Serious Upgrade, and I Can't Stop Using It
- AI-First Software Development: Redefining How We Build Software
- AI Code Is Not Slop
- Why "Good Enough" Code Wins
- The Future of AI Coding: Beyond Tools to True Autonomous Development
- I Asked 399 Developers for Their One Wish. Here's What They Said
- Measuring Developer Performance (And Why AI Might Make It Worse)
- Personal AI Agents: The Complete Landscape
Engineering & Dev Performance
Open Source
Product & Growth
Sales & Outbound
Pitching & Fundraising
last 30 dayslast 90 days