A brief background of the important news
On April 16, Boston Dynamics released a farewell video for the hydraulic version of the Atlas robot. The next day, they released their latest all-electric Atlas humanoid robot.
One of the great features of the electric version of the Atlas is that the joints and movements are very flexible. In the demo video, Atlas stands up from his back and performs a series of bizarre counter-joint movements. When the Atlas is fully standing, its neck and waist can be rotated 180 degrees in place.
According to IEEE SPECTRUM, Boston Dynamics plans to launch the electric version of the Atlas as a product in the future, mainly in industry, logistics and other fields.
So, what industry trends can we spot from Boston Dynamics' electric version of the Atlas?
The following is Wang Yuquan's important news commentary:
Recently, Boston Dynamics released a new video, which is behind a new major overhaul of their new all-electric Atlas, which is their humanoid robot.
What's the change behind that? It's that they've finally decided to ditch the hydraulic route in favor of the electric route.
Pay attention to the layman to see the bustle, and the insider to see the doorway. To the uninitiated, this all-electric humanoid robot can perform all sorts of unbelievable movements.
For example, a person's legs can't be bent like it, as if they were discounted, but they can miraculously stand up, including its head can also be turned 180 degrees, which makes people feel incredible.
From an insider's point of view, we should read out two major signals, both of which are very important for the development of the humanoid robot industry.
The first signal is that the future development of robots may really be all-electric. It turned out that Boston Dynamics had always insisted on taking the hydraulic route, and even because of hydraulics, it was once almost in a desperate situation.
Prior to this, Boston Dynamics relied on the support of the U.S. military for many years. But then the U.S. military decided to end its support for Boston Dynamics, because they felt that although the Boston Dynamics hydraulic route performed very well, the hydraulic power was too noisy and not suitable for the military, so the U.S. military stopped sponsoring it.
So Boston Dynamics had to commit to Hyundai Motor Company in South Korea, which shows how much pressure there was at that time. But under that pressure, they are still taking the hydraulic route, which shows that the hydraulic route is really difficult for them to change, why change it now?
I think it is very likely to be related to the comprehensive development of electric vehicles now, because when electric vehicles are fully developed, many technologies such as motors and electronic controls have also made great progress, which can be reused in the robot industry.
Including why we say why Tesla also makes humanoid robots, this is not accidental. Because they do have many advantages, in addition to the advantages of artificial intelligence accumulated with autonomous driving, the industrialization of electric systems is also a big advantage.
So you will see that in the future, it should be said that the electrification of the robot industry should become the mainstream, which is a doorway that an expert should see.
What's the second doorway? It's even more interesting, the electric humanoid robot made by Boston Dynamics can make all kinds of unbelievable and awkward movements, which just shows that humanoid robots should not be mainstream in the future.
Why do you say that? This seems to be a contradiction, people make a humanoid robot with ability to surpass humans, but you say that humanoid robots are not the mainstream?
The reason is actually very simple, why do we grow up like this today, we have no choice. There is a theory in evolution called Kluge, which simply means that evolution is continuous, and you can't overturn the historical accumulation of previous evolutions and start all over again.
That's why there is something in the human eye called a blind spot. It is because in the evolutionary process, at the beginning, our photoreceptor nerve penetrates from the front to the back of the photoreceptor cell, when the number of photoreceptor cells is small, the photoreceptor nerve will not block the photoreceptor cell, but when there are more and more sensory nerves, it always has to pass from one place to the back. Then there is no photoreceptor nerve in this place, and it becomes a blind spot.
Why can't you change it? Actually, it's very easy to change, you put this photoreceptor in the front, and put the sensory nerve in the back, isn't it over? But because the design didn't create a substantial obstacle at the beginning, it was carried over, and when the obstacle was formed later, it couldn't be pulled back, so evolution could only make superimposed progress on the original basis.
It's the same with our brain, our brain stem is the reptilian brain, the most primitive brain, and then the cerebellum is added to it, and now the brain is added, you can't throw away all the original things, just superimpose them on top, this is the law of evolution.
However, we all know that the evolution of machines can get rid of this route, and it can be designed from scratch, and there is no need to superimpose them like people, so learning people has become a chicken rib.
Boston Dynamics has given us a good example that we don't have to learn from people, we have to grow into arms, and then we have to grow into wrists, with 5 fingers on each hand.
We humans have a limited range of motion for each joint, but if it were a robot's joint, it might do 360-degree rotational motion. At this time, if 360-degree rotation is a better way of movement, but we humans can't evolve, then you don't have to design like a human, then it's not rigid, then isn't it limiting yourself?
Moreover, if it is not necessary to design the movement to resemble a human, then there is certainly no need to design the robot's body to resemble a human. Therefore, in order to make a humanoid robot, your actions have broken through the limitations of people, and you have to use human form, which is an inherent contradiction.
Then we think that the original purpose of the humanoid robot is good, and we hope that through the humanoid form it can do all the things that people do, but it doesn't have to look like a human.
In fact, Boston Dynamics has proven this from the opposite side, its joints do not necessarily move like people, they can surpass people, and of course the image does not necessarily look like people.
Some people will say that it is true that in real life, everything we deal with is prepared for people, why can't robots look like people?
The reason is simple, it could be in a better way. For example, you can handle many things at the same height as a human, but we humans still need to use ladders a lot of times. For the robot, if there is a lifting device, it can lift its upper limbs very high, and there is no need for a ladder.
So to put it simply, it is true that the robots of the future can solve all kinds of problems that people need to solve today, but they don't need to be like people.
Why not like people? There is a more detailed and comprehensive explanation, it just so happens that my last week's science and technology training camp course is about the development of the robot industry, I also made a detailed analysis, if you are interested in understanding, you are welcome to come to our science and technology training camp classroom.
The above is today's content, more detailed industry analysis and underlying logic I will share in the science and technology training camp. Welcome to pay attention to the global outlet WeChat account and sign up.
Wang Yuquan has to be heard, and we'll see you tomorrow.