The Future Issue
Our future is going to be fantastic. Humanoid robots are going to make manual labor a thing of the past. All the factories will be automated using humanoid robots. We will have all our deliveries done by robots. Our children will be handled by humanoid robots. Our pets will be taken care of by humanoids. Humanoids robots, as tall as us, as physically capable as us, will be armed as our police force. They will do all of the heavy industry, all of the light industry, all retail jobs, all medical services. They will be present, optimizing our entire economy. This will create a difficult to imagine amount of economic surplus, forcing our monkey brains to work to conceive of such abundance.
We never had this ability before, to create perfect physical substitutes for us. These robots will have dexterous hands, being double or triple jointed, capable of pivoting one way or another. Much in the same manner, our future can go one way or another. Instead of optimizing our lives, how will we know that these robots won’t harm us?
Currently, nearly all robots, and all humanoid robots, are made in China. This seems strange given how restrictive the American government has been of other Chinese-made technology. The United States has restricted China from selling its surveillance equipment here due to trust issues. Similarly, it prevented China from selling electric vehicles here due to unfair trade practices. How then can we trust China to manufacture humanoid robots, capable of being among us, with us, and between us, at all times ready to overpower us? Will we really trust or even want to have the hundreds of millions of omni-present, omni-potent humanoid robots being manufactured in China? While the answer is obviously no, our current trajectory has us sleep-walking into that outcome.
There are a few nightmare scenarios for having our robots manufactured in China. First, the parts made with chips, such as the actuators and the sensors, can be programmed to conduct espionage. The robots could record activities in commercial spaces, violating trade secrets, state secrets, and in the home, violating personal privacy. We solved this issue with phones manufactured in China - we can potentially solve it for robots too. The difference here is that humanoids can affect the physical world whereas phones were solely communication devices. Which leads to the second fear, the fear of autonomous physical agents negatively interacting with our economy. They could steal resources, destroy property, or even harm people.
Given these possible outcomes, it is reasonable to expect American startups to be focused on the issue of on-shoring the humanoid robotics supply chain. Yet, it is the opposite. Of all the companies developing humanoids, only Tesla has dedicated itself to vertically integrating the manufacturing of robots in the United States. The other companies are developing their software while relying exclusively on Chinese-made robots. Why? First, humanoid robotics are still considered an emerging technology that will need a few more years to be finished. The startups building the brain for these robots are only focused on solving the software issue, believing the hardware to be already solved. Second, these companies assume that it’s fine to manufacture the robots in China if the control software, the brain, is created by American companies. The firmware cannot harm the operator of the robot. Third, it is a matter of convenience and speed. Sourcing parts is a nightmare, especially in the opaque supply chain world. Chinese companies ship both parts and complete humanoids ready for usage by the robotics companies.
The Potential Solutions
Based on where each part of the stack of the humanoid robot is created, we can create a useful categorization of different availability and safety profiles for these robots to be used in the United States.
First, humanoid robots with all parts created in the United States. This means that software and hardware stack is entirely within the United States. This outcome would be the most expensive as it would require creating an entire robotics supply chain within the United States. The company taking this approach is Tesla, which has had success with vertically integrating electric vehicle production in the United States. It would mean the production of chips, which has the most complicated supply chain. This route is possible but very unlikely unless Tesla succeeds or the United States government heavily subsidizes it like it does the United States military supply chain. Even if it does happen, this route will be the most expensive until the robots begin to automate the supply chain.
Second, humanoid robots with all parts created outside the United States. This outcome would mean effectively no control over safety of these robots, both from foreign-control and from autonomous AI issues. There is absolutely zero chance that such robots will be permitted by the United States government. Thus, the United States would have no humanoid robots. This seems like the most unlikely route unless the American companies completely abandon software for robotics.
Third, humanoid robots with software created in the United States and hardware created in China. This seems like the most likely outcome as it is the current effective strategy of most robotic companies in the United States. This gives limited safety control. The brain would be programmed by the American companies but the firmware would still remain a source of safety concern. Foreign control of the robots could not be ruled out. Currently the government seems content to permit these robots but their effectiveness and number are quite insignificant. When that changes, the government will ban these humanoid robots to prevent any chance of Chinese control over millions of humanoids within the United States.
Fourth, humanoids with both software and firm-ware parts created in the United States. This outcome would require that the firm-ware parts are manufactured in the United States with enough scale that they are affordable. Such an outcome is possible, and certainly feasible given that it requires the manufacturing of only a few select parts. However, such an approach gives the highest level of safety control while avoiding some costs associated with manufacturing all the hardware.
All these outcomes are contingent on the choices of the United States government and the robotic companies. Many were not aware of the state of the field, ready to sleep-walk into the most likely outcome. We believe in a prosperous future. To achieve it, we believe that the economy will require humanoid robots. To have them, we will need all companies to begin to actively shift as much of their supply chain to the United States as possible as early as possible. This technology is too potent to be excluded from the United States. And make no mistake, unless the critical firmware-parts are manufactured in the United States, there will be no humanoid robots. There will be no prosperous future.