Wharton professor and AI researcher Ethan Mollick used Claude Code running Opus 4.8 to produce 'The Veil of History,' an interactive data visualization that models where a randomly sampled person would land across all ~117 billion humans who have ever lived. The tool draws on Population Reference Bureau birth totals, Maddison Project regional shares, and HYDE estimates, then propagates uncertainty through 4,000 Monte Carlo draws — yielding, for example, that 81% of all humans were born before 1650 and a typical random draw resolves to a pre-modern subsistence farmer in Asia.
Mollick explicitly framed the build as a multi-domain capability test for AI: research, statistics, code, and interactive design in a single session. The finished artifact includes a time-scrubbed choropleth map, probabilistic 'draw a life' sampling, and a transparent methodology section — work that would ordinarily require a data journalist plus a developer. For CTOs evaluating agentic coding tools, it is a concrete benchmark of what a single Claude Code session can scaffold end-to-end without human co-piloting.