From the supply side, IT giants are well aware that as long as third parties use their platforms, they will leave data on the platforms, and this data will be a big gold mine in the AI era. From the demand side, those application companies can utilize the AI cloud services provided by the big companies to enhance their competitiveness. So it's a win-win cooperation for both the providers of AI services and the companies that use them.
The second trend is that new and old AI companies will engage in white-hot competition around smart entrances. In the Internet era, almost every entrance will cause a scramble, and the successor of each entrance scramble will turn into the next giant. For example, in the Web2.0 era, the rise of interactive social networking formed a social portal, creating Facebook and Tencent. The era of artificial intelligence is no exception, and after the fierce competition, the next huge super enterprise will inevitably be born.
In the era of artificial intelligence, the entrance of services is mainly natural language voice interaction, natural language processing will be the main mode of human-computer interaction, who can make the machine more understanding of human language, who can potentially win. Artificial intelligence will fully encircle all aspects of our lives, which means that the competition for the entrance of artificial intelligence will be even more fierce, and the earlier you participate, the more users you get, the greater the chance of victory.
The third trend is that AI will occupy the living room, and voice interaction will become a mainstream TV application. The concept of smart home has been shouting for many years, Bill Gates made a clear description in the "Road to the Future" published in 1995, but such a description has not become a fact so far, the core reason lies in the lack of user demand.
Now, as TV screen sizes increase and video content explodes, the traditional remote control is becoming increasingly difficult to meet people's needs for TV use, and voice-based intelligent search and intelligent interaction are rapidly emerging, with natural language interaction becoming the standard way to manipulate a TV set. Traditional TV manufacturers such as Samsung, video content providers such as Netflix and Amazon, and IT vendors such as Microsoft and Google have all launched or will launch their own TV-centered voice interaction products, and the smart home is becoming a reality from the TV set.
The fourth trend is that smart toys will be one of the first successful application areas for AI. If AI is used to diagnose a medical condition or build a self-driving car, people often can't tolerate it making mistakes. But unlike toys, people have a high tolerance for toys to make mistakes. So, today, when AI has not been able to surpass human intelligence, smart toys are exactly the entry point that can make AI gain initial market acceptance.
In fact, there are already a lot of successful cases, for example, at the end of 2016, Cozmo, a smart toy that can play games with people, was sold out of stock. 2017 will have a large number of smart toys on the market. China's toy manufacturing industry is strong, but there is no brand advantage, if China's toy companies and high-tech companies to cooperate, with their own strong manufacturing strength and marketing strength, the trend of launching their own brands, will be a huge business opportunity.
The fifth trend is that commercial robots will play a huge potential in specific business scenarios. These robots are often made for a specific scenario, such as shopping guide, front desk, delivery, so the degree of effectiveness is obvious, in a small area is also easier to promote.
Many powerful robotics companies are developing general-purpose robots that can be used in a variety of business sectors, and they can adapt to different environments, but the downside is that they are more expensive, such as Japan's SoftBank pepper robots, which cost about 200,000 yuan each, not including follow-up service fees and insurance costs. The specialized robots are more focused on targeted problem solving, but because of their limited capabilities, they can easily be imitated by competitors or be annexed.
In 2017, we will see professional robots in many business halls with certain characteristics. In the future, the hotspot will shift from specialized robots to general-purpose robots, and general-purpose robots will rise rapidly.
The sixth trend is that assisted driving will become the first large-scale application of AI. Autopilot has always been the hottest application in the field of artificial intelligence, but due to the previous fatal accidents caused by Tesla's autopilot function, and Google has given up its own production of self-driving cars and switched to cooperating with mature car manufacturers, all these facts show that there is still a long way to go to fully realize the autopilot.
Meanwhile, the major car companies are making efforts on self-driving, more and more cars will be equipped with a certain amount of self-driving ability, that is, in the case of a driver, on the highway, or in the city at slow speeds to achieve self-driving.
The seventh trend is that AI applications will rapidly spread across application areas, and application startups backed by big data will have the best opportunities. Although the competition in the field of artificial intelligence is very fierce, especially on the underlying technology, don't forget that the real winner will not arise in the technology layer or platform layer, but in the application layer. For example, the rise of the mobile Internet, the biggest winner is not the communications operators, but the construction of the social network empire of Facebook.
And applications for innovative enterprises, artificial intelligence is precisely their real opportunity. Especially with IT giants going open source, there is an opportunity for innovative companies to capitalize on these resources and become the next hegemon. For innovators, it is imperative to go to the enemy's weakest link to find a breakthrough, and the most worthwhile place to start is where data is plentiful, but applications are barren.
The eighth trend is that GPUs will be further and more widely used in smart chips.GPUs, or what we call graphics processors, have always been the dominant hardware processors for AI applications, and they are rapidly expanding their market share in AI fields such as image and speech recognition and drones.