Robots on the move
Now that we are almost there with having real AI at our disposal — or so we are told — and judging by the fact that both OpenAI and Anthropic are filing for IPO getting them close to the $1 trillion mark, what’s the next frontier?
Self driving cars? Google’s Waymo is clocking miles, but its cars have a hard time distinguishing flooded roads from wet asphalt. All while Tesla’s fleet still needs the extra hands and eyes to get around the block. It’s promised to be ready in 2018 and every year after. But these are technical details and eventually they will be resolved. That’s the hope.
The next thing on the horizon are the robots. The army of soulless machines doing our chores like dishes, folding laundry or going to the store to pick up things for us. Funny things though. I wrote about something similar at the beginning of 2025. That’s when we were told that robots are almost here.
The funny part is that out of all the tasks we could come up with, these are the most trivial ones — dirty dishes and laundry. Similar to driving, which seems also so simple, dishes and laundry present a problem which the robots can’t manage.
A few challenges to consider.
The hand. The amazing thing which humans have and is subject to intense development. There are many companies in this space like Sanctuary AI or LinkerBot. It is not just the movement of the hand and the fingers, but the coordination between both hands and recognizing the object to handle properly. I am sure that with your closed eyes you can recognize objects and you know how to handle them.
Let’s say the hand problem is solved. How do you handle the dishes?
Unlike driving which has a set of rules and visual clues along the road, there is no manual on how to do dishes. Yes, we have the standard plates, cutlery and cups, but what about bowls, pots and pans. And the fine china left to you by grandparents which under no circumstances can be broken. How do you account and train for the endless variety of objects?
There are now companies which are trying to entice people with money and asking them to record their daily activities around the kitchen. From cleaning the dishes to cooking, people are recording videos and uploading them for further analysis. A reporter, in the name of research, tried to do that for a week and at the end earned $21.55. The positive side effect of that exercise was a very clean place with very clean dishes.
Now imagine that somehow the hand design is solved and the robot can do the dishes. Would you expect that the robot can do it perfectly, 100% all the time? Or would you be fine if 20% of the dishes would stay dirty. You might remember my post about a mushroom picking robot where the company realized that only 100% of mushrooms being picked makes economic sense. Anything else, and it didn’t work. And that was a robot designed for only one, very specific task — picking one kind of mushroom.
While a robot — a specialized hand on rails — can pick a mushroom, Nvidia is investing in a project which would see a robot 6 foot (182 cm) tall and 150 lb (68kg) heavy to learn how to do the dishes and laundry and everything else. And not to forget that Tesla is getting ready to produce Optimus in the soon to be finished factory, where — when in full production — it will produce millions of them.
In other words, we are many, many years away from robots in our homes.
While reading various articles on this topic, I found an interesting quotefrom a founder of one of the robotics companies: “People don’t really care whether they are unemployed; what they care about is whether they receive relief payments or welfare benefits.”
That line of thought is similar to all builders of AI — replace people with technology.
The good news is that in order for this to happen — to quote another founder— “I want every person on the planet to be recording themselves doing the dishes. That’s going to make a robot so that you never have to do the dishes ever again.”
The recurrent pattern? To avoid training the robots to take over the dishes and our lives, I’ll keep hiding them in our dishwasher. Faster and cheaper.