Markie Wagner is the founder of Poetic, a company building software that converts business processes into adaptive, deterministic code using frontier AI as a compiler rather than a runtime agent. Backed by Founders Fund, Kleiner Perkins, and OpenAI, she previously wrote the widely-circulated essay "Choose Good Quests." Her core thesis is that businesses mistake token spend for AI progress; the durable efficiency gain comes from using AI once to compile rules into code, not continuously as an execution engine.
Operating themes
- AI as compiler Argues that the unit economics of AI in production favor compilation to deterministic code over agent-based runtime execution.
- Token economics in business processes Quantifies the cost gap between runtime-agent and compiled-code approaches, using specific enterprise case studies.
- Business-process automation Focuses on high-volume, rule-bound processes where the compilation pattern delivers the greatest ROI.
Cards
- Using AI as a compiler converts expensive thinking into cheap deterministic doing, Using AI as a compiler converts expensive thinking into cheap deterministic doing [Tier B] (co-author with Packy McCormick)