The Test Set by Posit Sleeping Rats and Sociopathic Agents — with Phillip Cloud
Mar 9, 2026
Phillip Cloud, principal engineer at NVIDIA and Ibis lead with roots in early pandas work. He shares how an eye‑movement lab pulled him into open source. Conversations range from terminal user interfaces and VisiData to cold, task‑focused coding agents and their limits. He also riffs on file hierarchy grievances, NixOS praise, and a steadfast stance on pineapple‑on‑pizza.
AI Snips
Chapters
Transcript
Episode notes
From Eye Movement Lab To Pandas Core Team
- Phillip Cloud entered open source through an eye movement lab that used MATLAB, ported lab analysis to Python, and discovered pandas' column multi-indexes matched his neuroscience data needs.
- He contributed to pandas by tackling practical gaps like HTML/CSV readers and became an early core team member through those hands-on fixes.
Developer Productivity Shift Fueled By Modern Editors
- Developer productivity improved as editors and tooling (notably VS Code and plugin ecosystems) made work more accessible to many more developers.
- That UX ethos filtered into CLI tools, raising expectations that command-line tools must now offer polished TUIs and rich color/interaction.
Use VisiData For Fast Terminal Data Exploration
- Try terminal-first tools like VisiData when you want high-speed, keyboard-driven data exploration with both power keystrokes and optional menus.
- VisiData supports single-stroke composability plus clickable terminal menus so beginners can learn commands gradually.