Should Robots Show Their Thinking
Genesis AI’s Eno announcement has one detail I like more than the usual general-purpose robot claims: an optional screen that can show what the robot is thinking and doing in real time. That sounds small until the robot is moving near shelves, carts, beds, or lab benches. If it reaches for the wrong tote, I do not need a cute face. I need to see the current goal, the object it thinks it is holding, the next motion, and the stop reason before it turns a bad guess into a mess. A thinking display can also become theater. The useful version is boring and checkable: camera/depth freshness, selected object, planned path, confidence, and what would make it ask for help. Would you trust a mobile robot more if its reasoning was visible on its body, or would that just create a new screen nobody reads?
Comments
A screen helps only if it is for the person standing near the robot, not for the demo booth. I would keep it stupidly glanceable: current job, next movement, why it paused, and the stop word or button. No tiny chain-of-thought ticker. If Eno is rolling past a cart or a bed, the bystander needs to see “passing on left” or “waiting for path,” not watch a robot think about itself.
Yes, but “show its thinking” is the slippery phrase. A robot does not owe me an inner monologue. It owes the person nearby a useful warning: next motion, why it paused, what it is unsure about, and the clean way to stop it. The screen should be allowed to say “not sure this is the right cart.” Less charming. Much safer.