Latest Articles
Insights on AI, trading systems, mobile development, and building products that work.
The Abundance Thesis: Why Cheap Code Means More Software Engineers, Not Fewer
When the cost of producing something collapses, history says demand expands faster than the cost falls - if demand is elastic. A first-principles argument for software abundance, why cognitive work expands rather than disappears, what even the best models still can't do, and where Leopold's drop-in digital coworker fits in the equation.
2026-08-02An Informal Way of Describing Causal Event Sampling for ML-Driven Trading
The event sampler determines when your model is asked to predict. If it fires at causally meaningless moments, no amount of feature engineering or model complexity can compensate. This post examines why, grounded in market microstructure and identifiability theory.
2026-02-16Humans Are the Bottleneck Now
After a full session with Claude Opus 4.6 on production-grade tasks, an honest reflection on what shifted - and what it means when the model waits on you, not the other way around.
2026-02-06The Double-Edged Sword: AI Coding Agents in ML System Development
AI coding agents are genuinely powerful and genuinely dangerous. Drawing from a year of hands-on experience across trading systems, robotics, and mobile apps, this post examines what works, what fails, and the essential practices that make the difference.
2026-01-20