Here’s a video of the latest version of the humanoid robot Atlas.
Boston Dynamics has always been a leader in robotics, but there are many others not far behind it. Not only will robots like Atlas continue to improve, thanks to Chinese manufacturing they will get cheaper. UBTECH’s version of Atlas retails for $16,000. Some will quibble it’s not as good, but it soon will be. Not only that but in a few years’ time, many manufacturer’s robots will be more powerful than Atlas is today. Some Chinese versions will be even cheaper than UBTECH’s.
At some point, robots like these will be selling in their thousands, and then millions to do unskilled and semi-skilled work that now employs humans, the only question is how soon. At $16,000, and considering they can work 24/7, they will cost a small fraction to employ, versus even minimum wage jobs.
16k is insane. Even 32k is a super low for a bipedal, work ready robot imo
The UBTECH one is definitely not as advanced as the Atlas one. But I would expect, like everything electronic, China will eventually have commoditized versions of robots that are functionally almost as good as more expensive ones, but much cheaper.
https://www.techeblog.com/unitree-g1-humanoid-robot-mass-production/
No job is truly unskilled. Researchers have been trying to get robots to autonomously wipe tables and fold towels for years with only very limited succes. There is a reason why they only show the pick and place job here and nobody mentions the price for the training for a single task that’s not just picking things up.
Researchers have been trying to get robots to autonomously wipe tables and fold towels for years with only very limited success
Yes, this has been true up until now, but I think we are in a phase of rapid advancement. Look here at how DeepMind is using current LLM AI so that robots can train themselves - https://deepmind.google/discover/blog/shaping-the-future-of-advanced-robotics/
I would guess robots capable (perhaps messily at first) of general purpose skills like cleaning aren’t far off.