An open, hands-on
generative AI classroom.

Generative AI is collapsing the research crowd of collaborators, coders, and assistants into a single, AI-augmented researcher who can staff an entire lab with agents. I build that lab live: vibe coding tools from natural language, assembling teams of agents, wiring them into a shared knowledge graph that serves as a second brain, and orchestrating them into pipelines that run from question to result. But capability is not the same as quality, my research on human-AI collaboration shows these systems can raise individual work quality while narrowing the diversity of ideas a group produces, and that persuasive AI outputs can substitute compliance for the independent verification rigorous research demands. As the marginal cost of research approaches zero, the researcher's comparative advantage shifts from production to judgment, from generating results to knowing which expeditions to pursue.