AIGLIDE reportedly lifts robot teleoperation success from 0% to 70% on a plate-carrying task
A user describes GLIDE as using an LLM to predict where human robot operators will fail, then write and revise a filter between their commands and the robot.
TLDR
A user reports that GLIDE raised success from 0% to 70% on a task involving two independently controlled robot arms carrying a plate across a table. They describe an LLM that reads the task, predicts operator failures and writes a command filter, then revises it based on failures in the data. The user reports two other tasks improving from 10% and 0% to 90% each. They also say keeping the guardrail active at deployment raised a robot-control policy’s success from 0% to 60–70%.
