>>509664111>The robots will never break down and require repairsIt will, however repair cost can be factored in by simple replacement cost. In other words, instead of competing against a robot that costs 50k, simply move over and look at how competitive you are against a 100k robot.
>The robots will never have to charge a batteryThis is actually one of the tells - if a robot is using up enough energy to require frequent charges, that means that it's doing work. It's very much the same paradigm as task scheduling in operating systems - immediate, kneejerk intuition would be that you should minimize resource utilization to ensure it's efficient, but any resource that's not being utilized is actually wasted - it's not doing anything else.
>The robots will do the task at the same speed (if it takes 24 hours to do a 4 hour job, it's not cheaper)The problem is that now you're comparing on efficiency, and regardless of how slow it is today, it can only get faster.
>The robots will do the task at the same quality (if you lose 20% to QA, it's less efficient)see above. It's noteworthy that while human QA is practically irreplaceable for qualitative evaluations, mechanical QA is simply more accurate in every aspect. Mechanical QA (machine vision) is literally "eyeballing it" with a camera. No human QA can keep up with that kind of quantitative evaluation, human QA will likely shift to focus entirely on QA process evaluation improvement (telling the mechanical QA what to look for).
The real impediment to humanoid robot adaptation is that north american (and japanese) manufacturing still has a strong component of small to mid sized suppliers/factories where they have extreme specialties and such. If I were an economic planner, I'd be immediately worried that the big manufacturers would take their savings from humanoid labour, and immediately use it to cut costs by eliminating these suppliers through vertical integration. The rich get richer.