Figure AI ran a 30-hour humanoid robot livestream in San Jose, showing four robots (Bob, Frank, Gary, Rose) autonomously sorting packages at a $40B valuation.
Key Takeaways
Robots rotated shifts autonomously – pauses were AI self-resets, not teleoperation, per CEO Brett Adcock.
Human-form factor is the core bet: bipedal arms-and-legs robots fit existing warehouse aisles and workstations without facility redesign.
Real-world gap remains: crushed packages, torn labels, and high-speed throughput requirements still exceed current capability.
Figure positions the platform as updatable across verticals – warehouse today, construction and hospital settings next.
The 30-hour run is primarily a training data harvest and reliability baseline, not a commercial deployment signal.
Hacker News Comment Review
Commenters rejected the humanoid framing entirely: specialized non-humanoid robots have run factory lines for decades and outperform general-purpose bipedal machines on defined tasks.
The Shenzhen counterpoint was raised – fully automated facilities with specialized robots have operated for years, making a 30-hour demo look incremental at best.
Both comments converge on the same structural critique: humanoid form factor serves PR and investor narrative more than engineering efficiency.
