As a sci-fi fan and "Three-Body" fan, I have been looking forward to a domestic sci-fi movie since I watched "Dead Light on Coral Island" for the first time since I was a child. After waiting for more than 40 years, I finally got my wish in my lifetime.
I brushed "The Wandering Earth 2" on the second weekend. Teacher Ding believes that this sci-fi movie has surpassed 1 and the original work, and has become a sci-fi movie with great impression and depth. On the three-hour screen, a grand scientific, ethical, and philosophical depth is constructed. The film not only learns from the rigorous interpretation of Western aerospace science fiction based on real science and technology (rather than the magical world or unconstrained science fiction), but also carries the unique tenacity and great love of our ancient and vicissitudes of the oriental nation.
The special effects of this movie have reached the level of top sci-fi movies. And let's not talk about the magnificent scene special effects of the movie itself, but in terms of plot and script, it can also rank among the first-class or even top three. It is a literary and artistic work that can withstand the nitpicking of professors of science and technology and senior programmers, especially the easter eggs at the end of the film, which directly bring the narrative of time and space into the brain-burning concepts of parallel universes-I believe that Liu Cixin, the author of the original work, should be very satisfied with the script team . Today, let's talk about the computer technology in this movie.
1. Quantum computers
The protagonist throughout the movie should be the 550 series quantum computer. From the original 550A, the upgraded version of 550C, to 550W, finally formed a sense of autonomy.
Quantum computer is a new technology that has only been recognized by the public in the last 10 years. Although in the news reports, quantum computing technology has continuously made breakthroughs, but the current main application directions are still very limited compared to electronic computers.
The advantage of quantum computers over electronic computers is the concurrent superposition of multiple states. The parallelism of electronic computers is based on the scaled parallelism of 01 deterministic state. A gate circuit either represents 0 or 1 at a single moment. Assuming that all possibilities of a multi-dimensional vector are to be traversed, the intuition of contemporary programmers is to perform a for loop.
for (state i = 0; i < MAX_STAT; ++i)
{
evector v = evector::fromIndex(i);
//...
}
If v is very long and large, of course, it must be calculated in parallel:
#pragma omp parallel for
for (state i = 0; i < MAX_STAT; ++i)
{
evector v = evector::fromIndex(i);
//...
}
However, it should be noted here that parallel computing uses space for time. The scale of parallelism depends on the scale of the gate circuits involved in the operation, which increases linearly, and the benefits it brings are also linear. The fundamental reason here is that the gate circuit can only describe an exact vector v at the current moment, not all the values of v.
But for many high-dimensional problems, as the dimension increases, the calculation amount increases exponentially and factorially. Under the scale of O(n^m) problems and O(n!) problems, the benefits brought by expanding the scale and main frequency of electronic computers are not worth the explosion of scale.
A qubit is a physical register that can represent multiple states simultaneously. Before being observed, the state of this register is inherently superimposed. Therefore, if quantum computers have a general-purpose high-level language in the future, it is likely to look like this:
qvector i;
observe(calc(i));
Calculate and process all the states of a quantum vector i, and output (observe) the vectors that meet the conditions. Note that this calculation is applied to all states of the vector "at once", not a for loop. But as an electronic computer programmer, I can't figure out what kind of physical materials and devices are needed for this calculation (it must be an analog device, not a digital device). Because once the results are collected and output, it is equivalent to observing the quantum computing system. This observation would collapse Schrödinger's cats to a definite state (0,1). Therefore, there is a high probability that this language is digitally processed except for the output in the last step, and the rest of the process is analog processing. Boldly guess that the programming of quantum computers in the future should be "physical" programming, including microscopic 3D printing and many physical and chemical processes, similar to producing circuit boards and programming FPGAs.
In addition to editing programs and outputting results on electronic computers, the rest may be completed on dedicated hardware.
2. Artificial intelligence and digital life
In the movie, in order to deal with the solar helium flash crisis, there are actually many plans. The "Digital Life" project is one, hoping to upload human consciousness to a super quantum computer, so as to continue human civilization. In addition, MOSS also produces awareness, which is unplanned (at least unplanned by some people).
Letting computers generate self-awareness is neither the current development direction of artificial intelligence, nor is it difficult to have the hardware conditions in the short term. There are both ethical and technical factors here. The most obvious problem is that our current grasp of the physical structure and mathematical basis of the brain, life, and consciousness is still extremely limited and superficial.
Whether it is chatGPT or games, map recognition, and art creation applications, the semiconductor devices they train using massive data and massive calculations are still flat, linear, and deterministic electronic neurons, which should belong to the category of "bionics". Using "bionics" to construct neural networks is a plausible technical route, or a safe technical route.
The study of the brain itself is still far from "copying" consciousness. At least a few basic questions we couldn't answer.
- The exact model of the memory. The concept of hamburgers in our minds, recalling the appearance of steaks, what kind of physical and chemical structures and states are used to represent them in the brain. Some studies believe that short-term memory is related to bioelectricity, while others believe that long-term memory has certain molecular and microstructure components. But none of these studies can meet the standards of precision and rigor, let alone precise location and retrieval.
- A model of human brain thinking and decision making. Based on what kind of physical structure and information flow, we can plan our own activities, or just simulate such activities in the brain and get logical predictions?
- interface technology. How to accurately copy our brain information. Today's non-invasive and invasive brain-computer interfaces are still in the primitive stage. Through the processing of brain waves, a "black box" or slightly more advanced information-behavior mapping is obtained.
To achieve the acquisition, uploading, and operation of human brain consciousness, it depends not only on computer technology, but also on the development of basic physics, technology, and medicine. (Some people say that medicine is not science, which is one-sided. As long as this discipline uses objective methods to observe, count, understand the world and summarize knowledge, and guide the next step of action, it can be counted as science.)
3. Industrial automatic control
In the movie, the most impressive image is the door-type automatic machine. This automatic machine is a low-cost, versatile mobile aid. Judging from the performance of the plot, it is not a robot, but more like a tool that is remotely controlled by humans or controlled by AI.
This remote control device should be the technology closest to reality, and it can even be realized under the current technical conditions. Possess a certain degree of autonomous activity, through near-field wireless or wired operation (refer to the section on restarting the root server underwater). At the same time, its structural design in the movie is also very reasonable. After all, a tool facility does not necessarily look like a person, cost and multi-purpose are the most important. In the movie, some of the underwater door-type machines can only be remotely controlled by cables, which also conforms to the scientific setting of the underwater electromagnetic environment and the bandwidth of VR remote control.
As for the stupid puppy, it should be a bright spot. It should be a post-upload test for the consciousness of a natural dog.
From the perspective of art materials, all the scenes in the movie chose a conservative route, and there was no very peculiar levitation and minimalist design. After all, the setting for the outbreak of the Wandering Earth crisis is not far away.
But for the mechanical transmission, I still feel conservative. At least flexible materials and control technology are relatively mature in modern times.
4. Human-machine interface GUI
The man-machine interface in the movie, especially the GUI, is in line with the real technology and plot setting. A 3D projected GUI like in Avatar was not introduced into the movie.
The GUI in the movie is still an ordinary 2D, and the art and color scheme tend to be simple, practical, and eye-catching, which is very in line with the era setting and plot needs of the work. I am afraid that the software and art of human beings who responded to the crisis at that time were all for work, and high contrast, monochrome interface and simplicity were the mainstream.
postscript
Hats off again to the team behind The Wandering Earth 2, the best sci-fi movie I've ever seen. The cognition of technology and the prediction of technological development in the film are rigorous and scientific, the presentation of the work is grand and vigorous, full of respect for human beings, life, and culture, and full of nostalgia for history.