Andrej Karpathy rode in a near-perfect Waymo demo in 2014. It took a decade to become a paid product. That demo-to-product gap—not model architecture—is the binding constraint across self-driving, humanoid robots, and AI-powered education.

As AI agents gain tool access and long-horizon autonomy, the bottleneck shifts from model intelligence to governance—permissions, guardrails, monitoring, and liability. That's where job displacement becomes real.

YC's latest Light Cone episode argues that agents are becoming the primary selectors of developer tools, making documentation the new distribution channel. The companies optimizing for agent-parsable APIs and docs—like Resend and Supabase—are already seeing outsized growth, while legacy tools with human-first UX get skipped entirely.

The creator of a popular AI coding tool explains why they build for the model six months ahead—and why productivity measured by pull requests might be the 'simplest stupidest measure' of what's actually happening.