Tips on leveraging AI for empirical research

According to LessWrong, researcher Daniel Tan has developed a workflow for accelerating empirical research with AI. The approach: brainstorm an idea, spin up an AI agent to run experiments end-to-end, then review and refine the results. The keys to success are organizing research as atomic writeups—focused, self-contained experiments with a single headline result—and removing friction from reviewing and editing. Tan uses Claude Code and custom tools like a browser editor called Cowrite to streamline the process, allowing him to maintain multiple research loops in parallel.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton