Von Neumann robot of von Neumann machine

Klaus Knal is a professor at the center for geoengineering at Columbia University. Many years ago, he worked in Los Alamos National Laboratory. At that time, he and his friend Christopher Winter, a particle physicist at the University of Wisconsin, were having a beer and chatting. An idea came up: "How to solve the problem of global warming?"

They all know that this is no small matter. They want to convert carbon dioxide in the air into calcium carbonate-limestone, white marble and chalk, all of which are such things. The amount of work and resources needed to extract carbon dioxide from the air and convert it into solids is amazing. No one can count on investment or state funding, so they have to turn to other methods.

For example ... a fully automated transformation process. Suppose a machine. It can replicate itself. It can convert solar energy into needed electricity. It is easy to get the raw materials to make it. How's this?

That sounds great. The number of machines will multiply, from one to two, from two to four, and then to eight or sixteen. It keeps dividing like a cell, and endless solar energy will keep them multiplying until they stop.

Solar energy is not a problem. The raw materials needed by a machine are nothing more than common elements such as iron, copper, aluminum, silicon and carbon. These things are everywhere, just extract them. There is only one question: how to make the machine replicate itself?

Descartes was once stumped by this question, but now we have the answer. In the late 1940s, von Neumann had solved this problem preliminarily. At that time, he gave a series of speeches at the Hexon seminar in Pasadena, California. The core problem to be solved is "how does the machine replicate itself?"

Von Neumann believes that any self-replicating system should have two basic functions at the same time. First, it must be able to build the next generation with some constituent elements and structures consistent with itself; Secondly, it needs to be able to pass on its description to the next generation. He called these two parts "universal constructor" and "descriptor" respectively. Descriptors include a "universal machine" and descriptive information stored in a universal machine-readable medium.

In this way, as long as there are suitable raw materials, the general constructor can produce the next machine according to the instruction of the descriptor and transmit the described information to the new machine. Then, the new machine starts, and then enters the next cycle.

This idea was verified by an amazing discovery a few years later. 1953, Watson and Crick found that DNA completely met the two requirements put forward by von Neumann. Now, although we haven't designed a machine that can really replicate completely, self-replicating products using von Neumann's ideas do exist-computer virus is one of the most widely known.

Klaus La Knal and Christopher Venter intend to carry out their work according to von Neumann's idea. They named the project Akesson, a word borrowed from the Greek word "auxein", which means "growing up". They installed a high-temperature melting furnace for the original prototype to obtain the required metal raw materials and planned to throw them into the desert. There, adkerson can get a lot of raw materials and energy, and no one will bother. Although I didn't see the concept map, in my mind, adkerson should look like Kawara Tsutomu E. ...

For now, however, adkerson seems a bit too sophisticated. It is equivalent to stuffing a modern production line into a small box, which is still difficult to achieve with the current technical level. However, its creators are still very confident and hope to see this machine put into use as soon as possible. Think about its possible uses: producing enough electricity for the whole world, improving the global climate, and even changing the landscape of a large area, which is completely self-evident for hundreds of millions of solar machines.

This project did not win everyone's applause. Some people worry that this machine will destroy the ecological balance in desert areas and lead to unpredictable results. After all, we don't know enough about the planet we live in. It's hard to say what the result will be.

1995, Discovery magazine named adkerson as one of the "seven ideas that can change the world". That's true. For better or worse, constantly replicating machines will definitely change the world. Czech writer Karel Capec wrote a sci-fi three-act play called Russell Universal Robot at 1920. This is the first time people have heard the word robot. Nearly 90 years have passed, and robots and related products have already developed into a huge industry. Bill Gates, former chairman of Microsoft, wrote an article in Scientific American on June 5438+February, 2006, claiming that robots will become another technology product as popular as personal computers.

Although Bill Gates made many wrong predictions, this one seems more reliable. Household sanitary robots are not uncommon. Recently, some enterprises have introduced robot chefs. Children are holding coquettish robot dinosaurs, and the robot hostess is standing at the gate of the Science and Technology Museum. Some people even swear that by 2050, it will not be a problem to marry a robot.

In fact, it is not uncommon for machines to reproduce-such as our industrial assembly line. As you can see here at the moment, there are many invisible robots in the world busy in the factory. Japan has the largest number of industrial robots in the world, and half of them are in Japan. Those robots are different from what we see in comics-they just look like a mechanical arm, expressionless faces and handsome bodies. These robots weld and assemble mobile phones, computers and ... more and more robots appear on the production line day after day-but they are also unaware of themselves. This kind of robot looks much safer. Maybe soon, we can do the same thing at home.

In July 2008, at Cheltenham Science Festival, Adrian Bao Ye of Bath University and Vik Oliver of New Zealand announced a robot named "RepRap". This square machine looks like a shoe rack, and we can't see the shadow of the dream robot at all. However, it can replicate itself-though not completely.

RepRap can manufacture solid parts through computer instructions, and then the operator can assemble them manually. In fact, its core component is a three-dimensional printing nozzle, which is made of molten plastic or printed with molten low-melting alloy. Because not all parts can be made of plastic or this alloy, some parts must be made of other materials. This seems a little far from the self-replicating robot we imagined.

In fact, the 3D printer it uses is not new. The principle of a 3D printer is similar to that of a traditional inkjet printer, except that its nozzle can move horizontally and vertically to create a 3D image. Now there are some mature products, but the price has been high. RepRap has many advantages in this respect.

This is determined by its design purpose. RepRap aims to improve the present situation of those backward countries and regions. Baoye hopes that the large-scale application of this product can provide some manufacturing jobs for backward countries. You only need one RepRap and enough standard parts, and there can be countless. Then, you have your own production line.

This project is completely open and free. Anyone can download the relevant instruction manual, and the products made with it do not need to pay any royalties. Maybe it won't be long before we can make clothes hangers and slippers at home and make some RepRap for others.

Or there is another possibility. On May 1 1, 2005, scientists from Cornell University showed several squares. Each of these squares is 10 cm square and looks exactly the same. Stack four squares together and it becomes a very simple robot. It will look for a nearby square and spell it into a guy just like it. His designer said that there is a chip in each square, which stores the assembly instructions and controls the electromagnets on the surface of the square to complete various actions.

This kind of robot has a good prospect. We can imagine that a robot consists of dozens or hundreds of such basic units. When one of them is broken, it can be easily replaced. As long as the manufacturing process is simple enough, the cost of this robot will be very low and its uses will be almost endless. Maybe one day, this robot will produce its own basic units, and then we won't have to worry about anything. Imagine a little thing. It is about one seventh of the diameter of a hair. By the way, that's 1 nanometer. 1 nm is one billionth of 1 nm, and is only about 10 times the atomic diameter. On the scale of tens to hundreds of nanometers, people intend to make something. This is nanotechnology.

1959, Nobel Prize winner and physicist Feynman ("Stop it, Mr Feynman! This book is about this interesting physicist. He once gave a speech called "There is a lot of space at the bottom of matter". He predicted that mankind will be able to build materials with molecules and even atoms as basic raw materials in the most microscopic space. For example, we can arrange carbon atoms one by one into diamonds. After all, the world is made of atoms. In theory, nano-machines can make all objects.

Feynman's prediction soon came true. 199 1 year, a research and development team of IBM spelled out the word "IBM" with 35 xenon atoms on a nickel plate through a scanning tunneling microscope. Later, engineers made several nano-sized gears, scissors and propellers, but they never found a good motor to drive them.

Engineers need to learn from biologists if they want to make machines on such a small scale. Biologists have made a breakthrough in this respect. They found that there is a natural molecular motor in organisms, and all directional movements of organisms are related to it. Different types of molecular motors are also different, some walk on two legs, and some are divided into stator and rotor. With these things, those nano-scale components can be driven.

Some progress has been made in the control of molecular motors, and it will soon become a reality to drive nano-scale machines with molecular motors. But at the present technical level, the cost of producing such a machine is too high, and the best way is to use self-replication.

It will be easier to copy nano-robots. They can directly grasp the right molecules or even atoms to build a new self, or simply take advantage of the self-replication characteristics of DNA, from one to billions. Speed and cost are beyond the reach of other manufacturing processes. These little guys can work in the human body, cure diseases by killing viruses and germs, improve their physique by increasing oxygen supply, and even delay aging and treat cancer. These are all possible. Think about the Twilight Wars.

But at this stage, we can't expect this kind of machine to appear soon. Although such research results have appeared in academic journals and scientific journals recently, there is still a long way to go before they are really applied. With the new technology, it is often not just admiration. A self-replicating machine can bring unprecedented convenience, but its possible problems are also obvious: what if self-replication is out of control?

In 2004, a book named Kinematics Self-Replication Machine was published. On the cover of that book is a prairie full of rabbits. This cover is like a warning, reminding us of the terrible situation that infinite reproduction will lead to.

Machines are different from rabbits. Rabbits need food, excrement and their own territory. When there is too much reproduction, a large number of people will die because of the collapse of ecology, and finally restore balance. Machines don't need it. When its reproduction is out of control, we can only watch them devour everything, and the number will double every cycle.

Self-replicating machines have been concerned for a long time, and even a proper term "Gary Goo" has appeared, which is specially used to describe the scene that the world is swallowed up by constantly replicating nano-machines. 1986, Eric Drexler first put forward this term in his book Engines of Creation. This situation is terrible to think of, but it is unlikely to happen in fact.

Just as we install a switch on every machine, any machine that can copy must also have a control mechanism to stop copying. In our body, every time a normal cell divides, the telomeres of mitochondria shorten. When it breaks to a critical length, the cell will not divide, but will age and die. There are often such counter programs in computer software, especially some trial versions. Similarly, for self-replicating machines, similar techniques can be used to prevent them from being copied indefinitely.

But what if this mechanism fails? What if some guy with unkempt hair, wearing a white coat or an expensive three-piece suit deliberately released this infinite reproduction machine? What if one day the machine suddenly has self-awareness (I personally have no doubt that this day will come), becomes Skynet, and even turns our world into a matrix? These are still unknown. We can only hope on the wisdom of scientists, and can only say that whenever the world is on the verge of collapse, they have the ability to put the world back on track.

Scientists don't avoid danger because they believe they can control their own creation. Speaking of it, hasn't the whole history of human development gone through a similar path?

This article was reprinted at the Scientific Squirrel Meeting, and some changes were made.