Experience-driven notes on platform architecture, reliability, developer experience, frontend systems, and applied AI. The platform and reliability pieces come first; AI writing is kept when it contains a durable engineering idea rather than a trend summary.
Writing
AI infrastructure and agent systems
The Token Factory: Why AI’s Next Era Will Optimize for Throughput, Not Speed
A look at the token-factory thesis: lower-cost AI infrastructure optimized for useful output per dollar, not instant answers.
Applied AI and enterprise systems
Open Weights and the New Services Economy
Open weights widen the space for enterprises and service firms to build AI systems they can control, evaluate, and improve over time.
Agent systems and developer experience
Claude and the New Rules of Context Engineering
Why better interfaces, scoped context, and executable references matter more than ever-expanding instruction files.
Applied AI and reliability
Evals: The Missing Engineering Layer in AI Applications
A practical argument for grading both what an agent achieves and how it uses tools, permissions, and approvals.
AI standards and governance
A Possible TC39 Moment for AI
A standards-and-governance thought experiment for increasingly capable AI systems.
Platform and developer experience
Building Frontend Platforms Instead of Pages
A practical field guide to turning repeated frontend pain into paved roads without making the platform a tax.