Hao Cheng: founder of Xunlei and founding partner of Songhe Yuanwang Fund

Hao Ge shared six core issues of innovation at the 2117 product manager conference:

Next, let's take a look at what Hao Ge specifically said.

pure APP business, investors first ask you how to get customers, because everyone has only a few apps on their mobile phones.

PC Last Unicorn Company: Zhihu launched in 2119, and this is the latest PC unicorn.

in 2115, the competition of Internet was more intense than that of PC Internet in 2111.

the greatest value of the internet is to solve the problems of information asymmetry and information connection. For example, Taobao solves the information asymmetry between merchants and the connection between users and merchants.

but for many industries, information asymmetry and information link are not the pain points. For example, in the medical industry, doctors and ordinary people all over the country are linked together, which makes no sense. Because it does not improve efficiency, how can the Internet connect you? A doctor can only see one patient a day.

But AI solution is a matter of efficiency. For example, when a doctor watches a film, it can be solved through AI.

For example, Didi takes a taxi, which solves the problem of information link very well, but he still has an essential problem that he still has to rely on people to drive.

For example, in the catering industry, the transformation of the Internet is just skin deep, such as booking in advance and giving a discount coupon. But can the chef in the catering industry be replaced by the Internet?

So simply speaking, the transformation of artificial intelligence in the future is far greater than that of the Internet.

Therefore, the Internet and artificial intelligence solve different problems, so I suggest you pay great attention to this problem.

General technology has no successful business model. You can't sell money through SDK.

BAT knows that artificial intelligence is the next wave: they are not short of data, and all general technologies will be free in the future. for BAT, all these are wool on pigs, and this pig is cloud computing.

So entrepreneurs don't have the opportunity to be at the bottom

Industry +AI means: this industry existed before artificial intelligence, and through artificial intelligence, this industry can be changed.

There are entrepreneurial opportunities in these two categories, but:

The barriers in the AI+industry are very low, and the starting line is no different from the industry giants. You have no chance to understand this; And industry +AI: This is more friendly to entrepreneurship.

Every industry has its own depth: Hikvision: security company, with a market value of 311 billion yuan

Baidu's three major AI strategies at present: unmanned vehicles, DuerOS and Baidu's Feed stream. When Hao Ge once shared it with Baidu, Baidu people asked Hao Ge whether he should make use of his advantages in the technical service field, and Hao Ge said never:

The reasons are as follows: The depth of the security field is very deep, even if BAT wants to come in, there is no advantage. Therefore, industry +AI is a very barrier. Just like medical data is very difficult to get, even BAT has to be deducted from each department, and the team must have talents who really understand the industry.

can 99% cars get on the road? No, there are 1 accidents in 111 times! 99.99% is an accident every 11111 times! The difference between 99.99% and 99% is not 1.99, but 111 times.

The same is true for surgical robots, 99% of them stab people to death 1 times out of 111 times ...

And the research and development cycle of key applications is very long: there is a foreign company, which took 8 years from the beginning of the project to the sale of the first one, which is equivalent to at least 8 years of research and development cycle.

Therefore, the entrepreneurial teams in these fields are all "Gao Fushuai", with rich qualifications and background and super financing ability.

but not critical applications: there is no essential difference between the accuracy of 99% and 98%

For example, face recognition:

You can't achieve 99%, for example, you can't recognize it with a hat and mask.

If you can't recognize the face, you can still press the fingerprint. If the fingerprint doesn't work, you can still find the front desk.

this kind of application is OK when you reach a certain level.

what do you spell for non-critical applications?

the key application must be a big cow in the industry, and the non-key application is not necessary.

Only monopolists in the industrial chain can make money. Take a PC as an example: selling hard disks, monitors and whole machines doesn't make money. Who made money? Intel made money, and Microsoft made money, because they were monopolists, expanding to their own upstream and downstream.

Technology providers should try their best to do the whole stack: four in one

Only technology is available, but API is not available. Products must be commercialized, and commercialized data should be used to iterate their own technologies.

For example, robots have a certain amount of shipments in ToC, and there are only a few categories: sweeping the floor, drones and Amazon smart speakers.

there are more opportunities for tob:

in the artificial intelligence industry, technicians are the top leaders at present. Hao Ge wants to say that this is a temporary phenomenon. In the long run, it is still divided into two categories. The key application is that technicians are CEOs, and the non-critical CEO will be product managers or industry experts in the future.

to sum up, an AI product manager needs three abilities "PM+AI+X", and PM is the basic ability of a product manager; AI is the cognition of the red line of technology; X is an insight into the nature of the industry and understands the needs of users and the pain points of the industry.

That's the whole content sharing of Hao Ge and some of my explanations. It takes more time than expected to sort out this content, and there are still 7 articles to be sorted out. It is estimated that the holiday is so useless, and I hope it can be of all value to everyone.