What's Your Problem? Building a Robot People Actually Want
8 snips
Apr 30, 2026 Aaron Edsinger, co-founder and CEO of Hello Robot and former director of robotics at Google, builds practical home robots. He discusses why simpler, Roomba-like designs can be more useful than humanoids. He explains safety as a physics problem, the data limits for physical AI, teleoperation to mixed autonomy, pricing and real-world buyers, and modest design choices that enable useful deployments.
AI Snips
Chapters
Transcript
Episode notes
Keep Robots Simple To Ship Real Value
- Building simpler robots increases deployability and autonomy by reducing mechanical complexity and energy demands.
- Aaron Edsinger redesigned the human form into a Roomba-like base plus a telescoping arm to get 80% of humanoid tasks with far less cost and risk.
Sell Early To Fund Real Robot R&D
- Run a robotics company by selling real products early to fund R&D instead of burning investor capital on long development timelines.
- Hello Robot sells Stretch units (about 300 deployed) across institutions and countries to build a sustainable business.
How Stretch Restored Connection For Henry Evans
- Henry Evans, a quadriplegic, used Stretch to regain agency and reconnect with family through play and interaction.
- Henry controlled Stretch via eye gaze and teleoperation to play games with his granddaughter, who began calling him "Papa Wheelie."