On the evening of November 9th, Fu Sheng Mobile Chairman and CEO Fu Sheng pointed out at the 5th World Internet Conference that even on today's production line, it is difficult to solve the problem of equipment intelligence, and AI makes the machine the first time. It has the ability to perceive human beings and makes IoT terminals smarter.
“This is a huge watershed, making the concept of AIOT begin to appear”. Fu Sheng said that in the context of the industry and its own changes, Cheetah Mobile has launched its own artificial intelligence company, Orion Star, to develop speech recognition, visual recognition, The industry's only full-chain AI technology, including indoor navigation platforms.
“We will not make a single hardware in the future. We are committed to building an AI-enabled cloud operating system and working with developers to do more IoT devices,” Fu Sheng said.
The following is the full text of the speech:
First of all, let me take a moment to introduce the cheetah movement. Cheetah Mobile is a company that has experienced three leapfrogs. In 2010, it was formed by the merger of the cow image and Jinshan security that I founded.
The first wave we experienced at that time was the freeing of PC anti-virus software, achieving the first growth.
The second wave was on the mobile Internet, we met a strong opponent and decided to go out to sea. We are the first large-scale Chinese company to make users overseas. Today, Cheetah Mobile has more than 500 million monthly active users, 80% of which come from overseas. It is completely out of a different curve. We call it “The Flying Fairy”. So we are still the number one on the Google Tools leaderboard, and our games are downloaded nearly 2 billion times worldwide.
The third wave was after the listing in 2014. I realized at the time that the entire mobile Internet market was going to be exhausted and we needed to find the next opportunity. We have invested in funds in Silicon Valley and Israel, and we have seen opportunities for artificial intelligence.
I have some understanding and thinking about artificial intelligence. We have talked about the Internet of Things and intelligent hardware for a long time. In fact, before the emergence of AI or artificial intelligence, all intelligent hardware is pseudo-intelligent hardware. At that time, the intelligent hardware was essentially networking hardware, and you had to take out the phone to operate. About three or five years ago, the various smart hardware industries got up, but soon the industry declined.
I even want to divide the robot industry into two watersheds. Before the advent of AI, robots were essentially automated machines with no human attributes. We look at so many beautiful robots on the production line, which are actually produced through very primitive programming and automated equipment. Even today on the production line, it is difficult to solve the problem of intelligence.
But after the emergence of AI, for the first time, the machine has human perception. Our previous cameras saw a string of numbers such as 0101, but today the camera sees the human face; the voice that was seen before is the waveform. The voice we see now is text.
The machine began to have human perception, and the concept of AIoT began to appear. After seeing this opportunity, we went to the industry's only full-chain AI technology. Including speech recognition, visual recognition, navigation, robotic arms, microphone arrays, etc. Why do you want to do this?
This is to think from both inside and outside.
First, Cheetah Mobile has done too fast in the past. We just made a small tool and actually went public. Later, when Google’s policy did not allow us to make tools, and mobile phone manufacturers began to make their own tools, our days became very difficult. So I thought at the time that it was critical to have my own underlying technology.
Second, because the AI industry is still in its early stages. Many technologies are not perfect, or there is no industrial division of labor. However, many developers, taking a board and taking a so-called open platform, want to achieve a good speech recognition effect, which is actually very difficult. Take the smart speaker as an example. If you change the outer cover, the algorithm has to be adjusted because the reverberation is different. Just like adding an earplug to your ear, your hearing may drop dramatically, and it doesn't matter if your brain is smart. Therefore, the chain of the entire artificial intelligence industry is very new, and there is not enough division of labor.
Third, I realized that AI is not really an insurmountable black technology. Many people say that BAT can do AI, or people's accumulation is better. But today, open source code is far richer than any era. After many papers appeared on the other side of the ocean, Chinese engineers quickly realized it. So the benefit is that AI greatly reduces technical barriers. Moreover, AI has made the accumulation of many previous companies ineffective, and everyone is on the same starting line. We believe that this is a good opportunity to overtake a corner.
Therefore, we made up our minds to do all the technical work and make every effort to build a full-chain AI technology. The advantage of doing it all is that we can adjust a lot of details.
Today, I talk about artificial intelligence. Many people emphasize the so-called subversive intelligence. I have not seen the place to subvert people. But the core is that it is possible to achieve a little breakthrough. For example, when it can match the original 0101 signal to the real face and voice bytes, I think this opportunity is very huge. It will make many devices human-aware, doing a little work in a very vertical field. This little bit of work can bring about tremendous changes.
I hope that robots and speakers can cooperate with all walks of life. Our essence is not to make a hardware. Our essence is to be an AI-enabled cloud operating system, and do more with developers. Internet of Things devices. thank you all!