The Real Question Begins After the Trojan Horse Works
We treat “it worked” as the end of the story. The Odyssey treats it as the beginning.
When Cleverness Becomes Infrastructure
The smarter AI becomes, the less willing we should be to outsource the choice of what it serves.
Execution Is No Longer the Bottleneck—Turning Ideas into Systems Is
AI makes ideas easier to execute. It also moves the real bottleneck.
Rest Is Part of the AI Stack
The better you get at AI, the easier it is to overwork.
AI Native Systems
What I Learned From DeepSeek's Harness
A few days inside DeepSeek's open agent harness, pinned to one commit: four verified designs — the log the model cannot escape, the loop that is one config row, CI-enforced cache discipline, and the process OS behind 12,293 agent-written commits — and what builders should actually own.
August 16, 2026Read article
SignalDaily signals from my AI-native product research desk.
View all Today's takeaway
An Agentic IDE That Builds Itself; Alphabet's cash burn raises alarm for Big Tech as AI spending climbs; I built a watercolor Simulator based on real physics (V2)An Agentic IDE That Builds Itself; Alphabet's cash burn raises alarm for Big Tech as AI spending climbs; I built a watercolor Simulator based on real physics (V2)
AI Native Systems
August 2, 2026Should You Read AI-Generated Code? Ask What It Is Allowed to Change
A practical framework for deciding how deeply humans should review AI-generated code, based on the authority and consequences of the change.
AI Native Systems
July 6, 2026Expensive Tokens Won't Save Enterprise AI
Anthropic, OpenAI, AWS, and Microsoft are moving engineers closer to customers. FDE turns model capability into workflows the customer owns and can keep improving.
AI Native Systems
July 1, 2026The One-Person Project
AI coding is collapsing old handoff-heavy workflows into smaller ownership loops. The real challenge is keeping individual speed without losing team quality.
