Li Auto studies robot AI that retains experience for later tasks
On September 21, Li Auto researchers released ME-Brain-1.0, which stores experience externally for later decisions and actions.
They report 66.7% average success on six physical tasks, ten trials each. This does not demonstrate lifelong autonomous learning or a consumer launch.
Our analysis
Could a replacement robot inherit knowledge of your home?
We foresee value moving beyond the robot itself. Knowledge of dish locations, tidying routines and changing family habits could make a knowledgeable older machine harder to replace with an uninformed new one.
Portable memories and selective deletion could therefore matter. Giving users control over accumulated experience may shape durable services. These are our proposed next challenges, not announced product features.
How could everyday life change?
The following is a possible future based on this news.
A new robot could keep familiar household routines
Imagine choosing which routines to transfer to a replacement robot: children’s cups on this shelf, quiet tidying at night. A new machine could join the household without being taught everything again.
This is a scenario extending the memory idea. People would also need a way to review it so that mistaken experience is not inherited.
