Wang Xiang: physical-world data is ~6 orders of magnitude larger than language data
创业邦 · wechat · 2026-11-16
Wang Xiang (former Xiaomi president, now Alpine New Ventures) argues Physical AI adoption will be phased: service robots exist today, but core "robot brain" models are far from household-ready. Key point: while LLMs trained on hundreds of billions of tokens of internet text, physical-world data is estimated to be 6 orders of magnitude larger yet currently near zero — real-world capture (e.g. autonomous driving) and EMG signal decoding for action labeling may be breakthrough paths. He also stresses it's an ecosystem of brain + body (not necessarily humanoid), and calls for patience from both founders and investors.
More from AGI Musings
- antirez defends Huang staying CEO: investigate real wrongdoing before demanding heads — antirez · 2026-11-16
- After Cognitive Offloading: Humanity Revalues What Intellectual Work Means — Hub_Pli · 2026-11-16
- Safety as capability: why safer AI models get released sooner, and the case for open-sourcing safety evals — anpaure · 2026-11-16
- tszzl amplifies call: forecasters shocked by AI progress should be shocked by alignment progress too — tszzl · 2026-11-16
- Can we fix AI? Peter Bloem's new essay examines data, copyright and synthetic data paths — pbloemesquire · 2026-11-16
- Richard Dawkins asks Prospect: Can a machine be conscious, and how would we know? — connoraxiotes · 2026-11-16