Cross-task improvement reportedly not shown in five of seven reviewed robot evaluation protocols

A post describing “Who Improves the Robot? Toward Physical Recursive Self-Improvement” says five of seven robot evaluation protocols examined did not show improvement across tasks. The post says robots still depend on people to collect data, diagnose failures and choose which updates to keep. It also says the paper’s map of systems from 2022 to 2026 has no confirmed L5 case, the level for independently verified recursive gain.