Age of the Androids
What humanoid robots mean for the future of humanity
There is growing corporate interest in humanoid robots to replace human labor. Tesla’s recently unveiled Bumble C robot may mark a turning point in an industry that has thus far focused on specialized machines produced in limited quantities. Should Tesla succeed, what does a mass-produced humanoid robot mean for the future of humanity?
Of Legs and Brains
The story of humankind is often told in the frame of our advanced brains. Our cognitive capacity, indeed, is the cornerstone of our ability to use tools, farm, start fires…etc. Our brains enable us to build the foundations of modern civilization.
But this story leaves out another evolutionary attribute that also enabled all of the above; our bipedal ability to walk upright. This is equally important because in order to utilize the full capacity of our brains, we needed the capability to be mobile and manipulate objects simultaneously. This is something that this not possible when walking on all four limbs.
But ambulating on two feet is not easy. In robotics, bipedal machines remain one of the greatest challenges for engineers. So why bother? Why not simply focus on machines that use four legs, or even wheels? Certainly, these alternative approaches are cheaper and easier.
The easiest path isn't always the correct one. While four-legged and wheeled robots certainly have their place, if the goal is to design a multifunctional general-purpose machine that can serve as a drop-in replacement for humans, naturally, it must be bipedal. Why? Because we are. Our entire civilization, from the design of elevators, stairwells, to cars, buses, and even spacecraft, is built around having two legs with two arms.
Bipedal robots have been around since at least the 1990s. Most famously, Honda’s Asimo demonstrated the ability to walk and even jog, in the early 2000s. So what sets Tesla's approach apart from those that came before it? In a world: scale. Tesla’s vision far exceeds the more limited aspirations of Asimo and others like it.
First, a caveat, Tesla’s Bumble C robot remains far behind the likes of Asimo and Boston Dynamics, and this was painfully evident in the presentation that introduced it. Nonetheless, the fact that Tesla was able to develop a bipedal machine within a year, signifies significant future potential.
Tesla is already far along in the development of a second prototype, called “Optimus.” Optimus strives to be a general-purpose humanoid robot that will be produced at scale, leveraging the production and design methodology used on the Starship and Tesla assembly lines.
I dived into this topic more here, but this methodology is a five-step process that recognizes the (often forgotten) synergy between production and design. The philosophy focuses on rapid iteration that seeks to reduce the number of parts/processes while simultaneously improving the end product.
This is more difficult than it appears. Mass production is at least 1000 times more challenging when compared to building a small number of prototypes. Tesla ambitiously hopes that mass production and rapid iteration will lower production costs to under $20,000.00.
When combined with the latest AI, even a sale price of $40,000.00 could dramatically change the landscape of the economy. These androids could replace humans making deliveries, working on assembly lines, stocking shelves, serving drinks…etc.
All they would need is training from their human counterparts and their AI would learn as a human would. There would be no specialized machinery or coding. Instead, a drop-in replacement for human labor that doesn’t get sick, doesn’t need a day off, and doesn’t need a paycheck or health insurance. More than a few employers would find such a machine attractive.
In the Western world, robots, especially android/humanoids, are treated with heavy skepticism, and to a certain extent, fear. Much of this fear arises from a concern that they will usher in a future without jobs. This fear is likely misplaced.
Machines have long been replacing human jobs, from the spinning wheel to the cotton gin, to the robotic welding arms on auto assembly lines. In that time, however, the number of jobs available has steadily climbed and wages have risen. Paradoxically perhaps, more automation seems to create more jobs, it just creates different kinds of jobs.
Just as a larger population doesn’t automatically lend itself to resource scarcity, more machines don’t automatically lend themselves to job scarcity. Robots do not destroy human opportunity in the aggregate, instead, they increase opportunity by eliminating the need for manual labor and enabling greater specialization of skills.
There are some who argue that this time is different. In years past, machines merely supplemented human labor but didn’t have the physical facilities or AI to replace human labor. They argue that new humanoid robots may turn past experiences on its head by destroying more jobs than they create. This is possible but seems unlikely to me.
Regardless, some people in certain professions will inevitably lose their jobs to the humanoids. This transition is understandably difficult. This is why a strong, 21st-Century welfare system is needed to mitigate the perceived existential threat of innovation and minimize the socio-political pushback against progress. I have proposed some ideas here, such as a UBI or citizen dividend, and (crucially) delinking healthcare from employment as logical starting points.
Nothing to fear
Tesla’s announcement may have come at an auspicious time. In a world beset by labor shortages and facing a demographic cliff, there is a growing demand for a general-purpose android that can work the jobs we can no longer fill. While some might view the emergence of humanoids with apprehension, a future filled with robots is likely to be a positive development for most. But as with anything, policy and society must be ready if, and when, they arrive.