The ENPIRE research, covered by Import AI, points toward a future where robots could autonomously refine their own skills, but the path there is fraught with infrastructure challenges the report itself acknowledges. The real story may be less about imminent superintelligence and more about the immense engineering effort required to parallelize and instrument even simple robotic fleets. In our view, the most significant hurdle isn't the AI policy code, but the physical orchestration—the automatic resets, the failure recovery, and the management of idle robot time while agents 'think.' This suggests that for the foreseeable future, scaling physical AI will depend as much on mechanical and systems engineering breakthroughs as on algorithmic advances.