Robot dogs invade the campus
China's universities are secretly doing dogs???
Recently, many college students across the country have found that strange figures have begun to appear frequently in schools.
Some of them just wander around the campus when they have nothing to do:
Some will go to the playground of the school sports meeting as a conspicuous bag:
Some simply staged a "Cyber Burning Winter" at school:
Although today, mechanical dogs are nothing new, and this kind of crazy real dog occasionally appears on the city streets:
But when this "dog thing" really appeared on campus with such a swagger, many college students would still be stunned when they saw it:
I thought it was abstract enough to sneak an electric paper dog in the dormitory back then, but I didn't expect there to be a master......
(The "vicious dog" tied outside the dormitory door is my youth that I can't go back to, picture source: Internet)
Could it be that the mental state of contemporary college students is really so far ahead?
Not really.
Many of these mechanical dogs wandering around the campus are from the mechanical engineering schools of various schools to "let go", because many universities in China have been obsessed with engaging in mechanical dog projects during this time.
Mechanical dogs, also known as bionic quadruped robots, have been a key research topic in top universities in China in the past.
In March 1986, the National High-tech Research and Development Program (also known as the "863 Program") was promulgated, and intelligent robots were included as the world's cutting-edge technology.
With this as the background, since the late 80s of the last century, many universities, including Beihang University, Tsinghua University, and Shanghai Jiaotong University, have begun to research bionic quadruped robots.
(Image source: Internet)
It can be said that in the field of robot dog research and development, the academic community of mainland universities has accumulated decades of academic and experience.
However, because the research direction is mainly scientific research and exploration, rather than facing the consumer market, it has not attracted much attention from the outside world.
Born for research
As for the reasons why colleges and universities across the country suddenly "indulge in being a dog", there are two main reasons:
First, in the past few years, the domestic self-developed mechanical dog market has been developing rapidly, and it has basically achieved "open sales are not expensive", which has greatly reduced the research threshold of university research groups.
Second, the technical blessing brought by the outbreak of AI large models has greatly improved the level of intelligent interaction and machine learning of the robot dog, giving the robot dog more imagination space and application prospects.
(Image source: Unitree official website)
You must know that the mechanical dog is a natural integration of technological applications in many cutting-edge fields such as robotics, artificial intelligence and sensor technology, and is a complex of mechanical engineering, computer science, electronic engineering and other disciplines.
It plays an important role in promoting the development of cutting-edge technologies and cultivating cross-field innovative talents in colleges and universities.
(Image source: Tsinghua University official website)
Therefore, once the robot dog can achieve commercial mass production, it is only a matter of time before the campus is "invaded" in a big way.
However, what you may not know is that the current research on the application of mechanical dogs in domestic universities has shown certain results.
Electronic guide dogs have always been an important direction for the landing of mechanical dogs, and now it seems to be becoming a reality.
For example, this six-legged electronic guide dog developed by the team of Professor Gao Feng of the School of Mechanical and Power Engineering of Shanghai Jiao Tong University can realize the interactive and collaborative control of robots and blind people through the combination of natural language listening and speaking and force perception blind canes.
For another example, the R&D team of Northwestern Polytechnical University successfully deployed the offline large model to the low-computing power computing platform of the robot dog by using the latest large model quantization compression and accelerated inference technology.
In this way, even in a poor network environment, the robot dog can still maintain real-time intelligent response and human-computer interaction to cope with different working conditions.
More importantly, the combination of offline large models and mechanical dogs, so that the idea of "electronic guide dogs" is one step closer to reality, just imagine, a robot dog that does not need to be connected to the Internet can also realize intelligent human-computer interaction and action guidance, how many visually impaired people can help?
There are also some colleges and universities that are well versed in the truth that "the truth is only within the range of the cannon", and they directly fill the force value of the mechanical dog and save them from decades of detours.
For example, Nanjing University of Science and Technology recently demonstrated a sharpshooter mechanical dog, which can basically be used to shoot wherever it is, and hit every shot.
Earlier, this mechanical dog, which is said to have been jointly developed by Tsinghua University and a certain army department of the Western Theater, has been able to take the lead in charging on the battlefield, landing on CCTV, and walking to the pinnacle of dog life.
It seems that the same is a mechanical dog, but the fate between dogs and dogs cannot be generalized, some dogs are still in the playground as a conspicuous bag, and some dogs have been directly next level.
Compared with the various topics, in fact, there is a deeper connotation behind the "collective dog" of Chinese universities.
Because it completely shows how China's self-developed bionic robot industry uses the advantages of supply chain integration, industrial agglomeration, and policy support to quickly promote the commercial implementation of products.
In contrast, the dilemma facing Boston Dynamics, a pioneer in bionic robotics, may illustrate some problems.
Industry pioneers, trapped in technology
For a long time in the past, as long as mechanical dogs were mentioned, many people immediately thought of Boston Dynamics.
(Image source: Boston Dynamics official website)
In 2015, Boston Dynamics released a demo video of its quadruped robot, Spot, the first generation of its quadruped robot.
The Spot Robot Dog in the picture can not only cope with various road conditions freely, but also show a strong "sense of life" in its action posture.
In order to show Spot's product performance more vividly to the outside world, the staff in the video directly attacked Spot.
In the face of human violence, Spot's unyielding appearance of struggling and maintaining balance conquered all audiences.
Netizens around the world have said: Although Spot is not a human, you are a real dog.
This kick also made Boston Dynamics a world-famous "Internet celebrity technology company" overnight.
In the following years, Boston Dynamics updated some videos on YouTube to maintain the popularity, and accidentally made himself a YouTube millionaire fan tuba, and the video views could reach tens of millions casually......
However, although this mechanical dog does not need to eat dog food, it can't live on traffic alone.
Over the past decade, the problem of commercialization has been entangled with Boston Dynamics and the investors behind it.
Originally, Boston Dynamics' funder father was the U.S. military, and the research funding mainly came from the U.S. Defense Advanced Research Projects Agency, so it didn't need to think too much about marketization and product monetization.
Until 2013, Google spent $3 billion to acquire Boston Dynamics.
With no commercial prospects, Google sold Boston Dynamics in 2017 to Japan's SoftBank, which was betting heavily on the bionic robot track at the time.
Time flies, and in 2020, South Korea's Hyundai spent 880 million yuan to take 8% of the shares of Boston Dynamics from Softbank, becoming the new takeover.
At this time, the valuation of Boston Dynamics has fallen by more than two-thirds compared with 2013.
Behind this data change points to a cruel truth: under the gorgeous packaging and showmanship, Boston Dynamics has never been able to find a balance between technology and market, and is slowly losing the trust of investors.
During the period of frequent selling, under pressure from the management, Boston Dynamics also tried to pretend to "save itself".
In 2020, Boston Dynamics announced a limited sale of the Spot mini, priced at $74,500, equivalent to about 500,000 yuan.
Aside from the fact that it can be delivered wholesale to Meituan, the Spot mini really doesn't have much specific use.
Due to the unclear use case and the lack of down-to-earth pricing, Boston Dynamics has not released the final sales figures, but I guess it won't be too good.
(Perhaps, becoming a rich brother's barbecue grill is the best destination for Spot mini, picture source: Internet)
This is actually expected, after all, Boston Dynamics has taken the route of pure technology research and development from the beginning, rather than precise market application.
Mark Raelbot, the founder of the company with a Ph.D. from MIT, is the quintessential geeky boss who started his business more so that he could study AI and robot control as he pleased.
(Image source: Boston Dynamics official website)
In his opinion, talking too early about the commercialization of bionic robots is a "boring thing".
Therefore, over the years, all marketing activities of Boston Dynamics other than technology research and development are not so much to show muscles, but to appease the management so that it can continue to spend money on the team to engage in research and development.
It's just that even if Boston Dynamics is really determined to engage in commercialization, it is not an easy task.
At present, the main application scenarios of robot dogs are forestry, power inspection, fire rescue and other fields, but they are more used as a means to try and explore new technologies, rather than really saying what subversive experience can bring.
(Image source: Boston Dynamics official website)
In the face of high costs, the unanimous conclusion is:
Okay, there's nothing wrong with continuing to use manpower.
This is still the case in the B-end market, which is not short of money, let alone C-end consumers.
High-tech technology is burdened with the pressure of commercialization too early, and sometimes it is easy to fall into a kind of endless cycle:
The cost of the product is too high, so that it cannot be sold, resulting in the market scale cannot be improved, and the production cost continues to be high......
In the case of unclear demand, the mechanical dog wants to go from the laboratory to the market, and it is not enough to rely on the "far ahead" in technology.
Now, four years have passed since South Korea's Hyundai acquired Boston Dynamics, and this technology-first company has still not come up with a product that can be commercialized.
As is customary, the investors may already be looking for the next taker, but they don't know how much Boston Dynamics has left in its valuation this time.
The Chinese subvert the mechanical dog industry
So, what if, I mean, there were companies that could make a mechanical dog that performed similarly well right now, but could cost only one-50th the price?
The answer is, just kill the mad.
When you sell a mechanical dog for 500,000 yuan, I really have to think seriously about what this product can do.
However, if you only sell this thing for 10,000 yuan, then no problem, deal, let's make friends.
Judging from the various "stuck neck" news in the past few years, we have begun to have a consensus: as long as there are business prospects, no matter what the product, Chinese companies and engineers can make efforts to bring down the price.
Taking the latest consumer-level product GO 2 mechanical dog of the industry leader Unitree Technology as an example, the entry-level air is only sold for 9997 yuan, and it is more than this price to buy a real dog in a pet store.
(Image source: Unitree official website)
You know, this is not an electronic toy dog, but a mass-produced mechanical dog, which is not only equipped with an AI model, but also supports the expansion of various modules such as robotic arms and depth cameras.
In addition to Yushu, there are currently a number of start-up mechanical dog companies in China, such as Azure Intelligence and Yunshen Technology, which have launched a mechanical dog at the level of 10,000 yuan.
(Image source: Yunshen Technology)
In addition, large manufacturers such as Tencent and Xiaomi are also laying out their own bionic robot projects, which makes the business prospects of the robot dog look broader.
(Image source: Mi Mall)
Although the technical routes and product styles of these manufacturers are different, they all have one thing in common, that is, they are backed by China's complete industrial chain.
From the supply of upstream core components such as sensors, processors, and batteries, to the complete machine manufacturing process in the midstream, and then to the diversified market demand for downstream connections.
China's high degree of industrial agglomeration and policy and market drive have made the robot dog, which originally took at least ten years to achieve large-scale landing, complete the iterative process in just a few years.
This kind of thing can only happen in China.
Speaking of which, there is an even more magical thing: that is, even the United States, the home base of Boston Dynamics, cannot refuse the "law of true fragrance".
In September last year, the U.S. Marine Corps made a demonstration to show how they used a mechanical dog to carry a bazooka and launch it remotely.
Sharp-eyed netizens found that the mechanical dog used in the US military test was actually from China's Unitree Technology, and orders could be placed online with a price of only about $3,000.
For military equipment, this is simply the price of cabbage, it is too fragrant.
(Image source: Internet)
As a result, as soon as the report came out, the Americans were embarrassed all at once: I talked about boycotting, but I didn't expect that the inner ghost was actually myself?
The U.S. military quickly clarified that this is just a demonstration of the effect, and only Made in USA products will be used in the official procurement in the future.
Alas, if you have to get along with money, what can you do?
Of course, this is just a digression, but it also shows from the side how far China's self-developed mechanical dog products have developed.
Back to the industry itself, with the iterative upgrading of hardware, coupled with the blessing of AI large models, from a technical point of view, the mechanical dog seems to have stood at the key node of the transformation from toys to tools.
When the production cost is no longer the main factor restricting the landing of the robot dog, whether it is the large-scale application of the industry, or to assist the mentally handicapped, to provide emotional value, etc., it will be a matter of course.
Bringing technology out of the laboratory and into society may be the value of the modern industrial system, after all, for most ordinary people, it is enough for technology to make life better.
If you want to dig deeper, maybe who can answer:
Do mechanical dogs dream of electronic bones?
That's even better!