Presumably, this is a way to collect diverse training data for the robot to be trained on. Wash and fold as a service is valuable (to some people), and presumable the “extra steps” are offset with the in-home aspect of this.
Meanwhile, the ethical considerations are huge. Laborers are literally training their replacement, and probably at questionable wages. They’re also explicitly inviting someone into your home remotely, and that person can see and interact with your house. Feels like a privacy and safety risk. Additionally, it seems likely that this would be a literal Trojan horse to allow international labor to work within the US without dealing with actual immigration. Oh and just for good measure, it’s taking the jobs traditionally held by some of society’s least privileged and most desperate workers.
Anyways, if it actually works, I want one.
Edit: I feel compelled to note that apparently they’re hiring in Palo Alto for these roles, today.
Develop practices and training regimen at home office, pay well to attract quality talent, develop a positive reputation, lock users in with expensive hardware, outsource and offshore, enshittify aggressively.
Companies found out that hiring indians and teleoperate the “robot” is far cheaper than having an autonomy or AI algorithms with sensors on-board. Speaking of, all these food delivery “robots” were/are teleoperated as well over the internet as well.
I'd expect it to be a training session with admin privileges. Similar to a robot vacuum learning the layout of the house and mapping maybe? Just with added steps based on where the washing machine, detergent etc are located.
I don't think its a training session. Current AI models are pre-trained before deployment for inference. After the model is trained, they load it into the robots computer, and it runs inference with that model.
You can't train the model again because you don't have enough memory on the robot, but also even if you did its slow and consumes energy. You could have it train in some server but then every new skill would require you to pay the equivalent price for renting a bunch of GPUs for many hours.
What they can do is, for everyone, have a base model, and then improve it over time. Then, with software updates they can improve the set of skills the robot can handle out of the box.
But this is the problem with current AI systems, without a continuous learning capability, you're always limited to the "default skills". As soon as you have something out of the box for the robot to do, you end up needing Indians to learn it.
All of AI is flawed in this way. LLMs for instance have almost no continuous learning capability, that is why we don't have AGI yet. They can't learn new skills. Therefore, they can't adapt to new jobs they have not seen during training. They can't even play pokemon properly or any complex game for that matter, because games involve learning new skills during gameplay.
`The devil is in the details. Its completely different if one user has a custom trained model versus the whole user base shares a custom trained model. You have to overthink about these things carefully, otherwise you don't reach AGI.
It's most likely just a remote piloted session that's fed into the bucket for the robot to train on unfamiliar tasks/edge cases for known tasks. Falls in line with the true meaning of AI being Actually Indians.
Is this a humanoid robot that's controlled by someone in a call center remotely doing your laundry?
Putting aside ethical reservations about how much they are probably paying per task, that feels like wash and fold with extra steps.