The first house robot makes you wonder ...
The first robot vacuum cleaner, Electrolux Trilobite, was released around 2001 and cost around $1.500. Now, one may get a basic model in the range $100–$300. My experience is that despite the relatively easy things that it requires to do, almost twenty five years later, a descent model gets stuck almost every time when cleaning carpets, looses sometimes one of its two mop pads, gets stuck for other trivial reasons (frequent bumper errors). So, one has "prepare" the house in order for it to function properly.
When I first used it, it opened my appetite. I dreamed of a house servant, a humanoid, that I would be able to "teach it" to do things for me. Something with "open source software" that one might install on a humanoid robot, capable of moving around. That would ensure that I would be in control of it.
Today I read an article titled "1X NEO: The humanoid housekeeper still in training" in Medium. Well, things are not moving exactly in the direction one would imagine. You may see the release video here to get an idea of what the company is selling. Then look at the video that a Wall Street Journal reporter did for it.
Of course it is extremely expensive: $20,000. What however is concerning is that according to the above WSJ video "there may be a human operator behind the curtain pulling the robot strings" doing most of the things except really trivial ones.
As pointed out in the article, the real goal of this product launch isn’t to sell a household helper. It’s to collect training data, like Tesla did with the self driving cars. For LLMs to work, companies (like OpenAI) had to scan the entire internet. There is no equivalent with robotic data. Only now, things are much more serious. You have a machine inside your house, which has audio and optical sensors among other things. So, your personal data is exposed to an unprecedented level.
We may have two decades or more till these things become affordable. Cost however is not the only factor anymore. More important things are at stake. We need a solution that does not put our personal information in risk. One may ask "how are you going to do that?". Well, it is not easy but that is another topic.
