The Era of Compound Engineering — Kieran Klaassen, Every/Cora
Summary
One engineer shipped a full AI email client without writing code this year by building a "compound engineering" system that extracts judgment and taste into reusable knowledge loops—proving a single founder with AI leverage outperforms full teams using basic AI workflows.
Key Takeaways
- The bottleneck evolution: code → planning → deciding what to build → memory/knowledge systems. As each layer solved, focus shifted from execution to judgment, where AI leverage matters most.
- Implement the "human-AI sandwich" loop: humans provide judgment at brainstorming (start) and polish/taste (end), while AI handles planning/working/reviewing (middle). Extract learnings into systems after each cycle.
- One engineer building solo beats full teams because compound systems let you avoid repeating yourself—each brainstorm and decision gets stored so the next iteration includes that accumulated thinking.
- Allocate ~50% of effort to creating the feature itself, but spend critical time building the knowledge/memory system that captures taste and judgment—this is what compounds across iterations.
- Implementation is increasingly solved; the competitive advantage is judgment and taste—deciding when to use the model vs. when to turn your brain on, which you extract into reusable systems.
Related topics
Transcript Excerpt
[music] Hello. Hello everyone. Welcome. I want to start with saying I haven't written a single line of code this year. Um, maybe I haven't even looked at most of it yet. I do ship. Uh, I have a product I built that thousands of people use and trust with their email inbox, which is amazing. I'm actually proud of the code I ship and I'm proud of the product I ship. I've been doing this for two years and trying to extract my thinking and my taste into a system that compounds. And I'm going to share you how I do that. Lots of stuff you hear is like, "Oh, you should use this the factory dark factory do that blah blah blah all the new hip cool things." Uh, what I'm trying to do is not that today. I'm going to just show you how I work and hopefully share something that you can bring to your workf…
More from ai.engineer
- Agentic SDLC at Uber — Uday Kiran Medisetty & Adam Huda, Uber
- How I automate my own job at Hugging Face using agents — Niels Rogge, Hugging Face
- LLM Knowledge Bases: a practical guide — Ben Holmes, Warp
- The Dark Arts of Web Automation: Teaching Agents to Use Websites Like Humans — Corey Gallon, Rexmore
- Evolution of agentic surfaces — Gagan Bhat & Isabella Kai He, Anthropic