On November 18, during the third World Internet Conference, li Zhifei, the founder of an artificial intelligence company, went out and accepted a science and technology interview. Li Zhifei believes that artificial intelligence is currently facing a good opportunity for development. At the same time, the public’s understanding of what ARTIFICIAL intelligence can do and how fast it can develop in the future is still not very deep, and it is in the stage of idealism. So in the short term, it must be very clear that artificial intelligence in the short term is not comparable to human beings. In his opinion, we should make solid efforts to implement cloud interaction, natural language understanding and computer vision in various application scenarios, such as cars, homes and wearables.
Giiso Information, founded in 2013, is a leading technology provider in the field of “artificial intelligence + information” in China, with top technologies in big data mining, intelligent semantics, knowledge mapping and other fields. At the same time, its research and development products include editing robots, writing robots and other artificial intelligence products! With its strong technical strength, the company has received angel round investment at the beginning of its establishment, and received pre-A round investment of $5 million from GSR Venture Capital in August 2015.
In his opinion, investors and entrepreneurs should calm down and do not deceive themselves, the media, nor the country or policies. This is meaningless and will ultimately hurt themselves.
In the next three to five years, Li Zhifei believes that the development direction of ARTIFICIAL intelligence is very clear, mainly focusing on the following two areas:
First, products for consumer use. One category is the products of big companies, such as search, maps and video, which is to put the technology of ARTIFICIAL intelligence into it, and then improve its efficiency, make it more human, more personalized. The other is products for the masses, new intelligent hardware. Such as wearable devices, smart phones, VR, AR.
Second, products for the enterprise scenario. For example, for medical care, finance, real estate and other fields. It’s about how AI technology can be applied to these areas to improve efficiency or services. But in this process, AI technology may only account for 30%, 70% is the understanding of traditional technology and traditional business.
Specifically, in the next three or five years, we will focus on consumer hardware products and strive to improve product experience. At present, we already have smart watches, and we are cooperating with traditional manufacturers to try intelligent vehicle products. (processing)
The following is shorthand for the interview:
Reporter: How many times have you come to this world Internet Conference?
Li Zhifei: It’s my first time.
Reporter: how did you feel when you first came here?
Li Zhifei: This is the first time for me to attend this kind of official Internet conference, because in the past, it was only the Internet conference within the industry. I think it is quite different, for example, the participants, the form of the whole organization, and it is a very, very serious conference. I can also see that the whole government is embracing the way of the Internet. For example, what access control is now all face brushing, I think this is good
Reporter: Ask is an artificial intelligence company, a combination of hardware and software. What do you think about the current development trend of artificial intelligence?
Li Zhifei: I think the development of artificial intelligence should not be viewed in such a short period of time as this year and next year. This may be a further development. But there’s a big difference, for example, between the last two years and the last two years, for example, you might be talking about deep learning itself, but in the last two years you might be talking about its applications, like language recognition, like vision, like natural language processing, and after that, This ordinary people can feel what kind of artificial intelligence to bring these convenience.
I think this is a relatively basic technology in the past two years, but now it is relatively above, and it has slowly begun to be applied. And everyone’s understanding of artificial intelligence (ai) may also be more and more clear, may be early understand just feel it will soon come, immediately or would upset a lot of things, but in fact, we see today may everyone is really slowly, many companies such as we now whether large companies or small companies, are thinking about how to through these new hardware carrier, For example, the hardware in the car, or the hardware in the home, or the wearable device, to take this AI technology, and really put it into the actual product. I think it’s a very, very clear trend.
Reporter: You also attended the opening ceremony this morning, and listened to the speeches of some leaders and government officials. How do you view the current entrepreneurial environment, including the environment specific to artificial intelligence entrepreneurship, and whether there is a favorable development of artificial intelligence for our entrepreneurship?
Li Zhifei: Yes, first of all, I think artificial intelligence, whether it is research or entrepreneurship, must be facing a particularly good opportunity today, because all people begin to pay attention to it. Two examples, for example, Ma Yun talked about artificial intelligence today, and tomorrow there will be a mobile Internet forum. If you look, there are twelve panels in total, but eight of them have a title of artificial intelligence. He though the BBS BBS called mobile Internet, but it is clear that today’s mobile Internet all go to the direction of artificial intelligence development, so it is a trend of obvious, is that all people, whether it is a big company or government or small companies, are focusing on artificial intelligence, so it is very good on the one hand.
On the other hand, I think the public’s understanding of what ARTIFICIAL intelligence can do and how fast the future development can achieve what we want is not very deep, or the understanding is idealistic. For example, Ma Said this morning that today we are turning people into machines, and in the future we hope to turn machines into people. It is to move the machine closer to the direction of human beings. I think it is not appropriate for this stage in the short term. Because the machines that we have today, the algorithms that we have today, basically don’t have the capacity to be as smart as people, especially when it comes to this kind of entrepreneurship, this kind of thinking, this kind of understanding.
So at this time, we must be very clear in the short term, artificial intelligence in the short term can not be compared with people. And we should steadfastly take these today, such as cloud interaction, natural language understanding, computer vision, in which various application scenarios, such as automotive, home, such as wearable, steadfast put this thing done, and the user feel truly useful, not just in the media, or discuss in the film, and then in the past
Reporter: In fact, I know you have argued against super intelligence in many public occasions, and you think we are talking too soon. But this super AI, this thing of the future, is it the future, is it possible, or is it not the future at all, or is it just our imagination, that’s what I want to know.
Li Zhifei: In the past, I would have told you directly that I don’t think it can be done, because I don’t see a way to do it. But today, I’m a little bit more open, and I think there is, maybe there is. But I don’t think that’s really what we should be concerned about, because at any rate, if it does happen, it’s probably decades away. And what we should focus on today is to say, in the next two years, three years, five years, ten years, how we step by step to improve the technology of today’s AI, its intelligence really. It’s not that we’re looking at a very, very big vision 30 years from now, but we’re talking about that vision today, and we’re not talking about three years, five years, even one year.
So I think it’s true that most of the time we can’t predict the future, but we can’t deny it, it’s very likely that it might happen. But yes, I think today we still should, whether it’s going to happen or not, but I think the only thing that’s constant is that you have to really map out what you’re going to do one year, two years, three years.
Reporter: Or do you think not much about the future and want to do a good job in three to five years in advance?
Li Zhifei: It’s not that there aren’t too many, but I don’t think, say, five years from now, I don’t see a path where we can reach the so-called super artificial intelligence. Maybe five years from now, when we have a deeper understanding of the most basic problems of today, and many of the basic problems have been solved, I can probably see a very clear path to what the first step should be, what the second step should be, what the third step should be. So today I’m not going to think too much about the next 30 years, because I don’t have a path, but the next three years, five years, even two years, one year, I think it’s quite clear.
Reporter: So you think, but capital, if you go to the market to do this kind of heat speculation, it might do this.
Li Zhifei: Capital it won’t look for twenty years, I think it might also see it is five years, ten years, right, so I think this is actually, I also chat with a lot of investors have, in fact everyone will think that this is the future, but also all think today from existing startups, or from an existing product shape, really don’t see how we get to that we imagine the future. I think it is possible to need this, whether investors, or artificial intelligence entrepreneurs, is to calm down, do not deceive yourself, do not deceive the media, do not deceive the country, policy or what, I think this is not meaningful.
I think today is the day to think about what we can do in three or five years, and what we can do in the future if we succeed.
Reporter: What do you think we can do in three or five years? The second is “Ask” (08:01)
Li Zhifei: I think the application of AI in the next three or five years may have several major directions: First of all, the direction of the first big, 2 c, is to give consumers use products, there have two types, one kind is the company’s products, such as search, maps, video, I think it put the technology of artificial intelligence to the inside, and then to increase the efficiency of it, make it more humane, more personalized, I think it is inevitable that three to five years.
The second may be in for mass products, another may be I just mentioned this new form of hardware, such as wearable, such as cars, for example, such as smart phones, such as VR, AR, these things may be in the next three to five years, maybe you can see a big experience an ascending, This is for mass consumption.
Another category I think may be facing this kind of enterprise scenario, is the application of artificial intelligence, such as medical treatment, finance, real estate. I think how AI technology can be applied to that, and then make it more efficient, or make the service better, will happen here as well. But you can see on the inside may AI techniques in this, may be the last successful or not, it may be only 30%, and 70% said or your understanding of the traditional technology, the traditional understanding of the business, and even can get the data of traditional industry, and the ability to put your AI technology, deployment into their data center, I think this is particularly critical.
I think the future is probably in this area, and I think it’s very clear for the last three or five years.
Reporter: Go out and ask what you will be doing in three or five years?
Li Zhifei: “go out and ask” is actually mainly concerned about the second of the three I talked about in front of just now, why? We are a startup company, so we don’t have as big companies, have what a huge number of users of products, so we can’t go to enhance existing maps, or search engine, we are a startup company, we always want to pay attention to consumption and the product, from hardware algorithm software is yourself to do, but there are some categories, For vehicle products, for example, we may choose to cooperate with these traditional car manufacturers, because they have very good brands and channels, and they know the car industry very well.
We mainly provide AI technology, so we will do some exploration in each of these fields. However, for some ways of exploration, we may do some heavy work, such as making hardware, brand and algorithm by ourselves. In some hardware scenarios, we may be a little lighter, depending on what kind of partners we can find in the overall market. In fact, in these areas, we are indeed in some relatively big progress.
But in general, unlike most AI companies, we have always focused on consumer hardware products and new forms of hardware. I call these new hardware forms AI native devices, artificial intelligence native devices. Our phone, for example, is not an AI native device, because when we first made phones, no one cared about the AI. But today I think car, home, smart TV, smart watch, or AR, VR, I think AI may be a very, very important part of the first day.
Reporter: What are the possible areas of integration with traditional vendors?
Li Zhifei: THAT I just said, probably in the car. Cars are, I think, a consumer product, so it’s in the realm of what we want to do, but like finance or real estate or health care, those are things that we really don’t think about too much right now, because we think that’s a very different business
Reporter: But “Ask” has a lot of voice analysis skills, especially in the cloud. Does it say that I export my skills to finance or some other areas?
Li Zhifei: Now we don’t have a very active plan, but as a startup company, we are still very flexible. In fact, for us, the most important thing is that we can apply our AI technology to some scenarios and finally have an impact, which is what we want most. If we really can find a very good partner, and the AI technology, we put the AI technology into this scene, really can have a great influence, we will not rule out the future, but today is not really.
Reporter: One more question about capital. The “question” is about the C round of financing completed last year. Do you have any action on capital this year?
Li Zhifei: Actually, there are all kinds of investors looking for us all the time, but frankly speaking, many investors do not see clearly, because AI is not a very typical entrepreneurial industry. So we’ve been keeping, keeping in touch with them, looking. But now because of our last year, including the us, before we left a total of five times, basically still more abundant capital, and because of our main cost is in personnel, mainly employees some overhead, as well as the company’s basic operations, so basically we still have plenty of cash, so we didn’t say to the financing of hysteria.
But OK, if there is a particularly good opportunity to find a particularly suitable investor, we can develop quickly with these capital, then we certainly do not rule it out.
If we don’t plan to raise money this year, will we now be self-financing in smartwatches and our in-car products, compared to “ask” how much we spend on these two items so far?
Li Zhifei: These two, our smartwatches have been sold for more than a year, including we recently sold overseas, so the hardware itself, if you take it out separately, must be, at least, self-financing, and even some profit. With smart cars, we may, because we haven’t officially started selling them yet, we may start selling them next month, which is our “Ask the Magic Mirror”. Similarly, I think we will definitely reach, maybe in the next year, we will reach the state I said. There are other areas that we are still exploring.
However, due to the high cost of THE AI company, the personnel cost is also very high, and the work we do is relatively large, so the whole company is really not able to cover its own profits and losses. But I think that if we can maintain those hardware products, including some other cooperation we have in the future, maybe in the next year or even two years, for example, in the case of three or five hundred people, we may have a great chance to achieve self-financing. I think this is probably very, very rare in this kind of AI startup in the world.
Reporter: I want to ask a question in particular, since the introduction of smartwatches, smartwatches last year should be a free-for-all, everyone is pushing smartwatches. But may be happened some changes this year, “ask” should be in Chinese smart watch NO. 1, but now China’s overall smart watch market affected by some, including before I had some company in shenzhen of financing, financing problems, it is the aspect that is smart watch the market, what do you think of, Is it a direction of sustainable development in the future? What is the product positioning of the watch market in the future?
Li Zhifei: first of all, I think, smart watches the market must be on the rise, it is not in the fall, but may be early everyone on app watch such expectations are too high, does it reach you, it can immediately on behalf of the mobile phone, or to a great extent, the mobile phone market to rob to come over, this did not reach. But anyway the market is not going down, I think it’s going up, that’s the first one.
Second, I think the smartwatch market is really possible. In the next year or two, you can hardly see it becoming the most mainstream smart device. I think it’s really unrealistic. But “Go out and ask”, as an early entry, and has its own special core technology, in this particular core technology of the company, we in this, in fact, we are, as far as the smart watch, I think we can make a very good product.
You can imagine, for example, if you could sell 200,000, 300,000 units a year, or 500,000 units a year, the hardware itself, I think, would support a very good cash flow. So that’s number two. The smart watch positioning, product positioning I think or want you, in fact, we understand that the problem is how to solve the problem of the battery, of course, this is difficult, just need the second one is how to solve, let everybody feel smart watches is not just a mobile phone accessories, dispensable, so we made many attempts in this.
The first one is that we support 3G, so that there is independence. When you are in some scenes, like running, you don’t need to carry your mobile phone to receive messages, you can also make calls. I think this in itself is indeed we can see that this is a very, very right direction. The second is that we can combine many practical scenarios, for example we will immediately send a product, is to make mobile payments, and public transportation card, because our phone embedded NFC function, at the north of guangzhou, bus card, including some of the supermarket, you can brush directly handed the watch in the past, I think this is also a payment and ID system, which may make smart watches a necessity.
And the third thing is, I think this kind of exercise is healthy, we really do a good job with this algorithm, rather than saying that the step is always inaccurate, the heart rate is always inaccurate, or you have no follow-up action after the calculation. So this one we’re also doing in these sports health scenes, we’re doing this kind of refinement in some depth. In my opinion, independence, payment ID and sports health are just mentioned. If these three things can be gradually implemented, I think you may find that in the future, for many people, it will become a mere need.
Reporter: If you look at the “ask” market for smartwatches or smart car devices, you think it’s going to have a breakout point, and everyone is going to do it. Because we look at the watch is such an explosive force, but recently the smart rearview mirror has a downward trend, but also in the phase of elimination.
Li Zhifei: In fact, there are a lot of smart rearview mirrors on the market, but many of them have poor experience and poor reliability. This will inevitably lead to problems, that is, many users will use them once and never use them again. So when we entered the smart rearview mirror, it was similar to when we entered the smartwatch industry. When we made smartwatches two years ago, actually there were a lot of people making smartwatches in Shenzhen, but they did not do very well in many experiences.
I think this kind of voice interaction, or the whole, even the visual, for example, safe driving, a lot of AI applications, I think it’s going to be a big trend in cars, and I think it’s going to be something that can be made, that users feel is really valuable. So we’re taking that very long view.
Today, when we enter the field of smart rearview mirror, first of all, our mentality is that, first, we need to enter the scene now, rather than just saying, this thing can not be changed, the rear installation is also very difficult, never move. I think today has really come to a time, that is to say, due to a variety of information or education, we have a certain degree of acceptance of this, so I think first of all, now choose to enter the smart rearview mirror, if the user experience is good, I think it may have some explosion.
Reporter: There are two more questions. One is that creativity must be more important for ai companies, because we are all innovative enterprises. How does Go Out and Ask keep its creativity at the forefront of the industry in the future? As opposed to being outdone by our new startup?
Li Zhifei: I think creativity is up to you to a large extent. If you really look at the essence of things, it’s what kind of people are in the company. For example, there are about 300 people in our company today, and 200 of them are all working on products, from algorithms, from hardware, from software, from interaction, from design. So we go all the time, and that piece of investment is very, very big. So I think the “go out and ask” question is not the innovation piece, at least for now. I think it is more possible to strengthen the commercialization, and even to some extent, maybe the whole environment is not very good, we have to fight our own way out, I think this is the most important.
Innovation, itself, I think we have been is a innovation of a company, including my personal, I if just for the sake of competition to gain some market share, in an already very mature market, this is not our interest, we hope to be able to use the AI technology, create a new user experience, a new category, So I think innovation in the future we should still be a major investment direction of one of our major companies.
, of course, our company could indeed today with one of five people, or a small company to more than ten people, may indeed exist again, we are in their eyes, or sometimes the speed, has turned out to be a big company, these things we are really need to go to the interior, both in terms of organizational structure, or from the staff incentives, or from the company planning, It’s a lot of work.
Reporter: The last question is how to make businesses work better. That’s the market’s problem for me. That’s their problem.
Li Zhifei: Well?
Reporter: How do you make what you do better as a company, people seem to be saying, if you want to, because I know you’ve been trying to make Ask a company like Google that is creative and comfortable to work for. Now I emphasize the comfort of the company’s work, as well as the company’s environment, how to make the work better, what kind of environment I want to have in a company.
Giiso information, founded in 2013, is the first domestic high-tech enterprise focusing on the research and development of intelligent information processing technology and the development and operation of core software for writing robots. At the beginning of its establishment, the company received angel round investment, and in August 2015, GSR Venture Capital received $5 million pre-A round of investment.
Li Zhifei: I think how to make the company can actually work better, or more efficient, or be able to achieve even greater success in business, in fact, the most important, must first be a person, really can still when you are one hundred people to recruit to the good person, when you are two hundred people, or the same recruit good man, and not to say that the quality is falling fast, I think that’s number one. Today we “go out to ask” or maintain a relatively high standard, that is, the people recruited in all aspects may be very excellent.
Is the second in the thing itself, I feel or do you want to do is to make people think it is very exciting, we just mentioned this innovation, or represents the future of science and technology, very technical content, the transfer of values is one of the very positive values, whether on business or product, or technology, so I think the people and things, If they are all very positive, then I think the business or the entire organizational structure can be straightened out slowly.