Robot Demo Authenticity
Robot demo authenticity is the trust boundary Yao Song / 姚颂 draws in 173: 对话姚颂:深鉴、东方空间、再出发,「天才少年」十年后 for evaluating robotics and physical-AI companies. He says some industry demos use teleoperation video as if it were autonomous capability, and that Striding AI / 正行创新 should mark video speed倍率 and whether a demo is fully autonomous.
The concept extends Technical Demo Retention Gap into Embodied AI. A software demo can mislead about retention, but a robot demo can mislead about autonomy, latency, safety, dexterity, reliability, or operator involvement. Because physical robots are expensive and safety-sensitive, unclear demo provenance can distort financing, hiring, customer expectation, and data collection.
AI 不只比智商,WAIC 和 Kimi K3 透露了什么新竞争 adds a WAIC application of the same boundary. The hosts argue that embodied-intelligence displays should not be read as autonomous capability unless the viewer knows whether the action is teleoperated, scripted, single-behavior, or planned by the robot itself.
Key Claims
- Speed changes should be labeled because a sped-up robot video can hide latency or operational slowness.
- Autonomy status should be labeled because teleoperation, remote takeover, scripted playback, and full autonomy imply different model capability.
- Demo authenticity is not only ethics; it protects product planning and investor judgment from false milestones.
- The concept connects to Robot Active Use Metrics because repeated real use is a stronger signal than one polished video.
- Exhibition demos should be evaluated by autonomy status and task-planning evidence, not only by visual polish or number of companies on display.
Connections
- Striding AI / 正行创新 and Yao Song / 姚颂 — source company and speaker.
- Physical AI, Embodied AI, and Physical Intelligence System Stack — field context.
- Robot Teleoperation and Remote Takeover — legitimate supervision layer that should not be hidden.
- Technical Demo Retention Gap and Robot Active Use Metrics — adjacent evaluation concepts.
- WAIC and AI Demo Deployment Gap — exhibition setting where autonomy, deployment, and commercial proof can be conflated.