Embodied AI founders: models overrated, robots should work in supermarkets now
创业邦 · wechat · 2026-10-09
A DEMOCHINA 2026 panel with Wang Yu (Daimon Robotics), Mo Yilin (Lingyu Intelligence) and Qin Shentao (OriginFlow) on taking embodied AI out of the lab:
- No need to wait for "emergence": Lingyu uses an L2-style human-teleoperation model; robots already do supermarket food sampling, crab catching and drink mixing at roughly 70-80% of human efficiency, and real deployments generate interaction data labs can't collect.
- Touch beats vision for manipulation: Wang stresses contact force, displacement and deformation can't be learned from video alone; visuo-tactile sensors plus large-scale real human/machine data are the moat.
- Data infrastructure: OriginFlow captures non-invasive surface EMG fused with vision to estimate hand pose, force and contact state.
- What's overrated: "Models" — the real barriers are the body, system engineering and economic efficiency; data value depends on diversity and downstream success rate, not hours collected.
More from Companies & People
- India to Need 1.4M AI Professionals by 2026, 3.5x Demand-Supply Gap — nordicinst · 2026-10-10
- Yannic spent half his day with robotics-learning experts '4 standard deviations' smarter — yacinelearning · 2026-10-10
- a16z podcast: why boring SaaS companies are best positioned to win the AI game — andreisavu · 2026-10-10
- Apple ML hiring PhD interns for efficient multimodal models and video understanding research — CSProfKGD · 2026-10-10
- Plane claims 5M+ users, 5,000 paying business customers, 70% Fortune 500 reach — JosephJacks_ · 2026-10-10
- From ICLR volunteer to top CS PhD: Jeande's full-circle meeting with Sasha Rush — Jeande_d · 2026-10-10