Latest

In many enterprise domains, AI is most valuable for exploratory search that discovers better processes, then drives codification into deterministic workflows.

Why the hardest part of software is choosing a coherent frame for the domain, and how that framing shapes everything from primitives to how easily the system can evolve.

LLMs have reversed the traditional programming workflow - code now comes first and cheaply, while understanding follows. The locus of difficulty has shifted from construction to comprehension.

Entrepreneurship is the act of building an empirical theory about the world - discovering what people need, forming a coherent theory around that need, and testing it through a continual induction-deduction loop.

Instruction-following isn't enough; we need agents with judgment, context, and a sense of consequence.