Rohan Paul (@rohanpaul_ai)T2·78 pts
- MaxGlove captures direct pressure and tactile signals from human hands, solving the camera's inability to detect 'slip precursors' and 'force corrections'.
- Video-only policies must infer force, whereas glove data provides direct feedback, significantly improving grasp success rates (e.g., preventing glass from slipping or cracking).
- Maxinsights aims to accumulate 1 million hours of force data by 2027, on top of its existing 2+ million hours of real-world video footage.