Strange dichotomies are part of the humanoid robot world right now. Here are some examples:
- One of the reasons for the popularity of robots is that industrial companies cannot find enough people, yet the robot companies declare, “We are not trying to replace people.” I know that’s true, they did not set out to take jobs. However, the fabricating industry has wanted a solution for over a decade and robots are a key component.
- Despite being made of steel and hard plastic and wiring and cables and servos, the humanoid robots are the ones banged up after a year’s work, not us (plus we all come with a base-model feature called “healing”). We live longer than our metal and plastic friends too (for now, anyway).
- One of the main “for” reasons to develop humanoid robots is that the world is equipped for humans and if things are well-positioned for humans, they’re well-positioned for humanoids. Eventually, if the humanoids are successfully adopted, it will become an “against,” because we will need to equip the world as a robot needs it, not the way we need it, and the humanoids will specialize and need other accommodations.
- Humanoids are adept at adapting and will be adopted. We will trust them to be generalists and to learn based on past experiences. They will be the Swiss Army knife, and purpose-built, specific technologies will remain the best-of-breed.
We’re better at athletic pursuits too (except a machine called an Iron Byron hits the fairway every time, but it’s not humanoid). Here are a few examples:
Deadlift
- Human: 1,124 lbs.
- Humanoid: 220 lbs.
High jump
- Human: just over 8 feet
- Humanoid: 3.3 feet
Standing long jump
- Human: 12 feet 2 inches
- Humanoid: 4.6 feet
Running speed
- Human: 27.3 mph
- Humanoid: 9 mph
The main strength of the humanoids is that they are generalists. There are so many tasks inside today’s factories that the programmers will need to link the AI-enabled operating system to understand verb nomenclature like PICK, TRANSPORT, TRANSFER and give the humanoid several options to use with each. Is it up high? Then LIFT, BALANCE, LOWER will get the job done, as opposed to CROUCH, PULL, REVERSE STEP, GRAB, LOWER for a lower box or item. The humanoid will be able to figure out which of these verbs to translate into its motion, even if it’s a new task.
One of the main weaknesses is, you guessed it, humanoids are generalists. They won’t truly excel at any task—something else, a best-of-breed solution, would do a better job (more accurate, faster, etc.).
Morgan Stanley predicts that the humanoid robot market will be $5 trillion in 2050—a figure that includes repair, maintenance, and support—and other studies say the price of a humanoid will be down to $50,000 per unit in the U.S. Actually, some lower-end humanoids are going for less than that now. I’m not a humanoid fan, but I could see the prices coming down much further than that. Keep in mind this is only 25 years away, and in technology adoption, 25 years has a way of changing to either 10 or 50.
Much of what you see in the general press that shows humanoid robots or caninoid (dog-like) robots moving about a factory could well be a pilot program, not a true, permanent implementation. That’s okay, that’s where we are in humanoid evolution.
Maybe we will have over a billion humanoid robots by 2050, with 90% being used for industrial and commercial applications, as some studies suggest. Are we wasting resources by creating these robots in our own image will become apparent on the shop floor first? I foresee the humanoids of today breaking into several iterations in the future, each getting closer to its eventual specialized area rather than specialized task.