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 15, 2026
The Fed is watching AI now
In its May report, the Federal Reserve named AI one of the top risks to the financial system. A year ago, 9% of market contacts called AI a possible shock; this spring it was 50%. Big Tech is spending $725 billion on AI infrastructure in 2026, and the money has run ahead of the returns. When the Fed says 'bubble,' builders panic — but the bubble isn't your job to worry about. Buried in that same report is a risk that is yours: almost everyone now rents from the same handful of providers. Here's the part worth your attention.
- business
- architecture
June 15, 2026
The pilot was cheap. Production won't be.
Your AI pilot ran great and cost almost nothing. That number lied to you. When teams take an AI feature from pilot to production, infrastructure costs routinely run three to five times the original projection — and it's a big reason 95% of generative-AI pilots never turn into anything that shows up in the P&L. The pilot is cheap because it's small, watched, and runs on the easy cases. Production is none of those. Here's where the 3-to-5x hides, and how to price it before it ambushes you.
- architecture
- business
June 15, 2026
Your training data has a bill now
For years the working assumption in AI was simple: scrape whatever you can find and train on it. That assumption is dying in court. Music publishers are suing Anthropic for $3 billion, the AI music apps Suno and Udio have already settled and switched to licensed, paid models, and more rulings land this year. The free-data era is closing, and a price tag is going on the inputs. If you train or fine-tune on data, 'we'll just use whatever' is turning from a shortcut into a liability. Here's what changed and what to do about it.
- business
- architecture
June 10, 2026
“Managed agents” are convenient until you can’t leave
Google, Anthropic, and others are pushing the easiest pitch in AI: one API call and we'll run your whole agent — the sandbox, the tools, the memory, the state — on our infrastructure. It's genuinely convenient, and for a prototype it's great. But notice what you just handed over. A managed model API rents you the brain, which stays swappable. A managed agent rents you the entire nervous system of your product, and that's a much deeper hook. Convenience and lock-in are the same purchase here — and the bill comes later.
- business
- architecture
June 10, 2026
One model for everything is ending
Microsoft just shipped seven AI models at once — not one bigger brain, but a reasoning model, a coding model, a transcription model, a voice model, and more, each built for a single job. Meanwhile the frontier generalists keep getting more capable. Both things are true, and the gap between them is the point: the headline race is about one model doing everything, but the thing that actually works in production is a curated stack of specialists. Picking 'the best model' is the wrong question now.
- ai-native
- architecture
June 10, 2026
The risk was never the model. It's the system around it.
The most authoritative AI safety body in the world — a hundred-plus experts chaired by Yoshua Bengio, backed by thirty governments — just landed on a quietly deflating conclusion: the most pressing AI risks come less from the models themselves than from the systems companies build around them. Not the sci-fi misaligned superintelligence; the integrations, the permissions, the business processes a small error propagates through. For builders that's good news, because it means AI safety is mostly a job you can actually do.
- architecture
- security