Cheaper than illegal maids?
2 Likes
This is very exciting.
I’d like to see a video of it cleaning a bathroom – even with a human behind it with a joystick.
intercst
1 Like
Given the strength of robots I would like to see it clean my large braided rug which is hard for me to lift. I would like to see it scrub my shower from top to bottom in places I can’t reach easily. I would like to have it scrub the inside of my dishwasher and oven because bending and reaching on my knees isn’t as easy anymore.
The question is whether the robot is only capable of simple activities, like wiping a countertop, or the harder jobs also.
Wendy
4 Likes
If it has a camera, it can be used to spy on your home.
I’ve recently seen YTs linking it to China. I suspect domestic “watchers” are also present. As well as anyone any where with the skills.
With the rise of LLM AI, see Anthropic Mythos, the skills required to hack these camera enabled domestic appliances, are within every body’s reach.
.
..
ralph
1 Like
Worth noting perhaps that they are talking about their “one” test from a waiting list, it shows that the robot was there for 3 hours (we have two people for 3 hours), and that the robot is “delivered” and “picked up”, which means there are gonna be some human costs involved until that gets sorted out.
And I suspect the “size” of the house doesn’t matter - until it does. Else they will have robots tied up for days on end in mansions, roller skating rinks, and the like. Im gonna guess it’s more like “whatever it can do in 3 hours”, not “size of the house is irrelevant.”
1 Like
I mean, somewhat?
Again, this is all about data. Data, data, data. As we’ve discussed ad nauseum in the threads about humanoid robots, the major hurdle for IRL AI robots is that there doesn’t exist anything comparable to the Common Crawl for the real world. There’s no big dataset for training.
So the AI and robot companies have to create it on their own. They strap folks up in haptic suits and teleoperate a robot; or load them up in sensors and have them wipe the same counter over and over and over and over so that they have the data to train the robots.
And that’s what this is. These aren’t robots that are commercially ready to be autonomous maids. It’s a company that is looking to have homeowners defray the cost of teleoperation training by letting the data collection company teleoperate in their home, rather than in their own facility. The upside is that they get a little revenue to defray the cost of the data collection and they get to teleoperate in a vast array of environments. The downside is that their data collection is slower because the units have to move from location to location, so they’re not collecting then.
5 Likes
The teleoperation angle is something most people completely miss about these robot maid companies. You’re basically paying to host their data collection operation, not getting a finished product. Albaby nails it — real-world training data is the actual bottleneck here, not the hardware. For anyone tracking emerging tech investments and lead performance across sectors, Phonexa gives solid pipeline visibility. This space will be fascinating to watch over the next few years.
1 Like