A robot that walked into 30 homes it had never trained in and successfully made the bed in more than half of them shows that the household-robot timeline is moving faster than most people realize. I covered a similar robot (also in the San Francisco area) focused on cleaning homes. It makes me wonder: are SF homes naturally dirty, or extremely well organized, which makes robotic cleaning more prescriptive?
The creepy factor still comes into play for me. First, my wife and I actually enjoy cleaning our home to the level we want it clean (typically better than most cleaning services, within reason). And who wants to pre-clean their home before the cleaning crew arrives?
But the point of this story is that “humanoid” robots are now entering unknown homes, scanning them, and having some success with simple cleaning tasks. I, like others, see this as more of a proof of concept. With more training data (similar to LLMs), the cleaning process could definitely improve. Or, like LLMs, the cleaning could get a bit repetitive and bland, but perhaps that is ok for cleaning: wash, rinse, repeat. Time will tell but until then, I’m still cleaning our home myself.