The "little i robot" in Yuan Hui's eyes: fanatical dreams, Transformers and the metaverse future

dialogue

/Hi TechWalker/

Tech walkers

The 20-year entrepreneurial history of Yuan Hui, the "father of little i", tells us that AI is ultimately to serve the people.

The author | Zhou Ya

Pictured| Hu Jie

Yuan Hui has always had a dream in his heart, just like the science fiction movie "Transformers" described, to make intelligent robots human partners. 20 years ago, he plunged into the deep sea of AI and became the "father of small i robots", where he was a little closer to his dream.

"Transformers" as a metaphor, represents the most perfect artificial intelligence, but also a certain common desire of human scientific and technological civilization - human beings in the past 60 years of historical development of artificial intelligence (AI) poured too much imagination and practice: from the first computer Turing, people have successively invented a variety of intelligent machines, can play Go, can deliver, act as assistants, automatic driving and so on... This can't help but make people wonder whether the arrival of Transformers is not far away.

However, there is always a distance between reality and dreams. It seems that these colorful AI are always a little bit worse than the ultimate image in Hollywood blockbusters. Not to mention that some people criticize "artificial intelligence" to some extent "artificial mental retardation", and even there is a strange phenomenon in the industry - AI only makes money and does not make money, and even the most popular "AI Four Little Tigers" are also losing money.

Yuan Hui, the "person who came over", also experienced the same wind and rain. As the founder, chairman and CEO of Xiaoyi Group, in the past 20 years of entrepreneurship, Yuan Hui and his xiaoyi robot have spent a complete process from the Internet, to the mobile Internet, and then to the era of the meta-universe ahead, during which they have the glory of being wildly sought after by capital and the media, and have also experienced the difficulty and even the silence of the life hanging in the line. However, he still sticks to the AI track, only because he believes that AI will eventually open a new world, and human beings will eventually create that new species.

01

Technology investment: Challenging the limits of AI

Among all the AI companies in the world, the "small i robot" created by Yuan Hui himself can be called one of the earliest entrants, and it has also created many precedents.

Chatting with AI with wake-up words like "little x, little x" may have become commonplace. But in fact, "little i" is the originator of the mountain. As the world's first chatbot, "Little i" was born in 2004, when IT was also the popular 3G era of MSN, Yahoo Tong and other (instant messaging tools that have now withdrawn from the historical stage) - this is 7 years before the emergence of Apple's Siri.

Behind the record-breaking product is Yuan Hui's clear positioning of the company. Click on the official website of Xiao i Robot, there is such a eye-catching slogan: "Explore the business future of cognitive intelligence", Yuan Hui told me, "cognitive intelligence" is the goal that the company has been challenging.

Generally, artificial intelligence can be divided into three directions: motion intelligence, perceptual intelligence, and cognitive intelligence. "Motor intelligence" represents positioning, planning and control capabilities, such as the dazzling Boston Dynamics robot dog; "perceptual intelligence" is sensory ability, common face recognition, speech recognition, image recognition, etc., which is also the fastest part of the current application, such as health treasure, brush face payment; "cognitive intelligence" refers to "can actively understand and think" - if you use human metaphors, "motor intelligence" is like limbs, "perceptual intelligence" is like eyes, ears, nose and tongue, "cognitive intelligence" is equivalent to the brain.

The ideal form of cognitive intelligence is what Yuan Hui calls a "transformer" with "concepts, consciousness, and concepts.". But there is still a long way to go to achieve it, AI is both multidisciplinary, data-dependent, and ethical, and the challenges are beyond what humans have ever known about all technologies — "even to let humans re-understand the world and know themselves," Yuan said.

"In the next 50 years or even longer of the golden period of AI, cognitive intelligence is the most core part." But insiders, including Yuan Hui, know that this is the crown jewel that everyone wants and is not easy to obtain. Especially as a startup, the first thing to do is to solve its first obstacle - commercialization.

02

Business model: Another hurdle for entrepreneurship

On the road of "cognitive intelligence" business model mining, Yuan Hui and the company have undergone a major transformation, like a rebirth.

The first key node occurred at the beginning of the business. In 2004, the number of small i-robot users for the C-end crowd has exceeded 100 million, and even today, it is a good achievement. Through the IM tool portal, users can add "answering robots" that can chat with you about gas, music, and even stocks, a "general artificial intelligence" product that can talk about anything, which cannot be compared to the world - "everyone thinks this is the future of technology". For a time, capital pursuit, magnesium light focus, subway advertising debut, Yuan Hui also became the front page headline of the media.

Therefore, according to the inertial thinking of the traditional Internet, the first business model that Yuan Hui chose for the company was the traffic model of "first do users, and then make money". Relying on the continuous door-to-door financing, Yuan Hui has taken it to continue to expand the scale of users. "At that time, investment, and no business model, depending on whether you were popular or not, the number of users was not much." In this way, financing, being a user, refinancing, and being a user again seems to be a wave of progress, but it is actually involved in the vicious circle of traffic bubbles.

By 2006, the number of users of Xiaoyi had reached its peak, and the huge number of users brought huge costs in bandwidth, server loss, maintenance, and manpower. At the same time, the first few financings have basically burned out, and if you want to continue to develop, Yuan Hui must hand over a satisfactory return answer to the investor.

By 2006, the number of users of Xiaoyi had reached its peak, and the huge number of users brought huge costs in bandwidth, server loss, maintenance, and manpower, but the commercial returns did not appear at the same time. Xiaoi has tried the conventional monetization models of advertising, games, and e-commerce, and obtained income through the way of "diversion traffic and earning share", but these are still a bit thin compared to the huge cost expenditure, and it is difficult to offset the income and expenditure. Until the outbreak of the financial crisis in 2008, the retreat of capital became the last straw that overwhelmed the balance, and there were only two ways in front of Yuan Hui's company - either turn or die.

Looking back on the past, Yuan Hui frankly said that many entrepreneurs, including him, have mistakenly believed the two sentences that are popular in the Internet circle, "as long as there are more users, you can make money" and "(more than a billion Chinese) each user gives you 1 yuan, you have developed" - unfortunately, to this day, there are still many people selling this methodology.

At the company's lowest point, Kazuo Inamori's "Living Law" inspired him. When Kazuo Inamori started his second business, he didn't think about commercial profits at all, and only put how to meet people's needs first, and eventually he became another Fortune 500 company. If you can really meet the needs of others and do this well, your business model will naturally form. ”

Yuan Hui concluded that the failure of the toC business model is also rooted in the fact that Xiaoyi did not provide the toC services that users needed. The universal, omnipotent AI service, not only the small i at that time could not do it, even now, no one can still achieve it. Blindly pursuing short-term traffic and ignoring the actual needs of users is putting the cart before the horse.

01

"Everyone wants to build Rome, but Rome is not built in a day"

In 2011, Xiao i finally found the grip of transformation. The company's positioning has shifted from omnipotent "general artificial intelligence" to "narrow ai" focusing on specific scenarios, providing commercial solutions for B-end users.

Obviously, ten years ago to take an AI toB road, Xiao i is a forerunner, the road ahead has no taillights, and there is no successful copy of the model, how did Xiao I get up?

This has to go back to the earlier 2006, xiaoi robot accidentally came into contact with the first government customer: the Shanghai Municipal Science and Technology Commission, for which developed a 7x24 hours to receive consultation customer service robot "Mr. Hyde", effectively alleviating a large number of inquiries in the government window. This project did not make much profit for Xiaoi, but it laid the groundwork for Xiaoi's transformation road.

With the typical case of "Mr. Hyde", Xiaoi began to contact more B-end customers. After exploring countless scenes, Yuan Hui finally chose "intelligent customer service" as the main battlefield. After more than ten years of accumulation, Xiaoi's "intelligent customer service" market share accounts for up to 90%, and has won nearly 1,000 large customers in China, including the three major operators, the top 50 banks, technology giants (Huawei, Xiaomi, JD.com), SF and so on.

Why is the "intelligent customer service" industry ice-breaking? Yuan Hui joked that don't look at intelligent customer service "old and dirty", but until now it is the industry with the best AI monetization ability, and the user's application benefits are also excellent. For example, Xiaoi has helped the credit card center of CMB in this regard, and through the intelligent customer service service, the energy efficiency ratio has been increased by three times, and 99% of the questions can be answered by it. More critically, "intelligent customer service" has also been upgraded from the "call center" in the 1.0 era to the "contact center" in the 2.0 era, and the AI dialogue is more natural, integrating service and marketing - "once deeply cultivated, there will be a clean future." Yuan Hui stressed.

Starting from the "intelligent customer service" track, today's Xiaoyi has opened up more application scenarios in the AI+ industry and gained a firm foothold in the field of AI toB. Yuan Hui outlined that there are four typical application scenarios for Xiaoi at present, the first is "call center"; second, AI+ construction, helping the construction industry to review all-intelligent drawings, and even reach the level of domestic CAD; third, AI+RPA, to help enterprises realize robot process automation, with their own 1,000 large customers as endorsements; fourth, soft and hard integration, empowering community epidemic prevention, public security anti-fraud, medical equipment and other industries.

For example, at the moment of normalization of epidemic prevention, Xiaoi urgently launched two hardware products: "external call machine" and "automatic external defibrillator", the former in order to improve the efficiency of community epidemic prevention, the latter for heart, asthma and other emergencies. "These two hardware are just a small test knife, but they are just needed, and in the future, they will be like fire extinguishers that are standard in every building." Yuan Hui pointed out.

Looking back at the transformation process, like Asimov's "Three Laws of Robotics", Yuan Hui summed up "three principles of artificial intelligence commercialization":

1. From simple to complex. People's understanding of AI is a Transformer, which is a misunderstanding, and any new technology must gradually extend from "simple scenes" to "advanced scenes". If enterprises only stick to the highest goal and continue to invest capital and technology, sooner or later they will dry up their resources.

2. From virtual to physical. Some robots themselves look cool, but there is no brain just an empty shell, then just add a little intelligence to it. The summary of Xiao i is: customer-centric, market demand-oriented, with the most appropriate software technology to build, with the most appropriate hardware carrier to present.

Yuan Hui took the example of the small i-out call machine, which does not have high technical requirements, but it can really solve the problem. He reminded that technicians are generally reluctant to use unexcient algorithms, but AI is not only to solve the pyramid spire problem, but also to empower those "traditional industries that have a huge stock of upgrades" in the tower, which is the biggest value of the current AI era.

3. From B2C to B2B. B2C is the Rome that everyone wants to build, but Rome wasn't built in a day. The maturity of AI requires application scenarios and industry support, and finally it can achieve a trickle into the sea. "Giving up the C-end dream to the B-end, the whole process is very painful and tangled, but today it is an inevitable process."

Yuan Hui believes that with these three laws to support the commercialization of AI companies, in time, it will eventually wait for the arrival of Transformers.

04

Towards the Metaverse: Instead of All in, we Always in

So far, Xiaoi has gone through 18 years, completed the integration of "technology + business", and ushered in the third key node: the arrival of the metaverse era.

"Someone asked me if I wanted All in metaverse, and I said, Little i Robot is actually Always in Metacosm." Yuan Hui means that Xiao i's AI ability has long been prepared to meet the arrival of the meta-universe, which can be described as a spring breeze, and gives the "syllogism" as evidence:

1. The core of the metaverse is AI. In addition to human-to-human interaction in the metaverse, a larger number of behaviors are human-computer interactions, and the core of the latter is AI. At that time, the chatbot Xiao i, in the metaverse, could also incarnate as a "virtual person" to chat with people, the times evolved, and the essence did not change.

2. At the same time, AI is accelerating the arrival of the metaverse. Because the metaverse has a great feature, that is, its means of production, can be infinite supply, if we build a CBD in the north, it will take at least a year, but if in the metaverse, it may be generated in a second, which is a spiritual world of "human thoughts> AI things".

3. Therefore, The ability of Xiao i in the field of cognitive intelligence can also be replicated and applied to the metaverse world. In 2022, Xiaoi will lay out three scenes in the metaverse: virtual people, NFTs, and AR/VR - because these scenes are "the most easily perceived by the people", for example, the current software entrance is an App, and the entrance to the metaverse may be a virtual person. Speaking of this, Yuan Hui concluded that the true meaning of the metaverse, like artificial intelligence, is a process for human beings to "re-recognize themselves" in a world where virtual reality is born.

Twenty years is not short, and the product iteration and commercial evolution of The Small I have never stopped. However, behind the countless changes, there is one thing that cannot be separated from its origin, AI is embodied in the technical concept is "cognitive intelligence", embodied in the business philosophy, is a universal "serving the people". Yuan Hui said that from the day xiaoyi was born, there were five words next to the original logo - "serving the people".

Because, whether in reality or virtual, whether in the physical universe or in the metaverse, it is always not wrong to use AI to serve human society to a better future.

-

This article is originally produced by Technology Walker, please do not reprint without permission