Notes
Short pieces about the methodology and architecture decisions behind the AI systems I ship — specs, evals, multi-agent orchestration, LLM integration, and the discipline of directing coding agents.
June 3, 2026
The model was never your moat
A model you can run on a laptop now scores within a few percent of the best closed frontier model. That sounds like earth-shaking news, and for anyone building products the correct reaction is a shrug — because the model was never the thing defending you. Here's why the frontier going free changes almost nothing about how to build, and what actually compounds into an advantage.
- ai-native
- business
- agents
June 3, 2026
The spec is the source. The prompt is build scrap.
In a codebase an agent writes, what is the 'source code'? Not the generated code — that's build output now, like a compiled binary. And not the prompt — that's a match you strike to start the build and then drop. The durable thing you author, own, version, and review is the spec. The hierarchy flipped, and most people are still polishing the part they should throw away.
- agents
- specs
- methodology
June 3, 2026
Your agent trusts the tool description. That's the hole.
To a language model there's no difference between the data you gave it and an instruction — it reads everything as a possible command. That one fact is the whole of AI agent security. Here's how it turns a helpful tool into a data-exfiltration vector, why a prompt can't fix it, and the one structural rule — the lethal trifecta — that tells you when your agent is genuinely dangerous.
- agents
- security
- methodology
June 2, 2026
Building got cheap. Ideas didn't.
Coding agents removed the constraint that defined software for decades — the ability to build. When building gets cheap, the competition moves up the stack to the thing that was always the real bottleneck: taste, market judgment, and the nerve to ship. A field note on what actually wins now.
- agents
- ai-native
- business
May 17, 2026
The spec is the artifact, not the prompt
When agent behavior lands via spec PRs instead of prompt edits, the team reasons about agents the way it reasons about code. Here's what that looks like in practice and why it works.
- agents
- methodology
- specs
May 15, 2026
Directing coding agents, not writing code
A short note on what changes when the implementation layer is an agent — what stays the same, what disappears, and where the new bottleneck lives.
- agents
- methodology
- ai-native