nativewolf
Active Member
That is exactly the question. The real costs are more though, to Amazon that robot replaces a worker with a 10% absentee rate, a high rate of workers injury claim, and probably $22/hour so a real costs of $35 an hour probably. Of course Robots break down and have issues. Lots of people ignore those challenges and sometimes the uptime is far less with robots than good employees. Then you need scale, you want repetitive activity that can be trained but over time the NNs might make the robot more competitive.It may be possible with the current bleeding edge technology to create a robot, which could fetch materials from a warehouse. But what would that robot cost? My WAG is that at least 100 kUSD. Why would you rather not pay someone $ 10/hour to do that?
Amazon has spent billions automating factories, they continue to look. They just haven't, to date, had someone in a silly suit dancing on stage.
I think having a robot bot would be great for many menial task and for helping monitor elderly and assist elderly with household chores, etc. I don't see a bot as necessarily a great replacement for many tasks. Tactile issues, which are a big deal, are dismissed by many posters that don't understand how hard that is. In this regard, Tesla/EM statements on FSD are coming back to the fore. It's always been next year. Well as far as I know noone has really solved the critical tactile challenges with humanoid robots. It doesn't matter what the NN can instruct if the mechanical bot can't perform. Humanoid bodies are amazing, it's going to be much harder to replace than some here believe.