Feng Jianxiong took a look at his thoughts and talked eloquently: "I think it's quite easy to prevent the imitation of Gu Yongjiang." First of all, we N station will do a good job of confidentiality in the future, and try to dilute the advantages brought by this innovation, so that Gu Yongjiang feels that 'even if N station catches up, it is only because we took the money of the Ahri department, and we are more willing to burn it'.

I know General Gu Yong very well, and I have formed a strategic cooperation with him for more than two years. He is a typical capital operation entrepreneur, CFO background, not very sensitive to new technologies, only sensitive to the R&D investment output ratio of new technologies. Even if he finally finds out that my layout is excellent, it will not be possible to copy it at all if he doesn't spend half a year investigating and making up his mind, and spending another half a year doing algorithms and building a team.

And for artificial intelligence things to work, the feeding of big data is very crucial. It can't be done overnight. Therefore, compared with Youku, we have a 2-year lead in artificial intelligence distribution, and I think it is not difficult at all. This gap is enough to cover the valuation difference I want today. ”

Jackma and Cai Chongxin glanced at each other and nodded secretly, knowing that Feng Jianxiong's answer to this angle was correct.

Jack Ma didn't bother to compare, so he chose to grasp the big and let go of the small, and asked the next key: "So, how are you going to prevent Zhang Yiming from copying your back road? I think that for you, he is a more dangerous opponent than Gu Yong on this issue." ”

……

How to prevent Zhang Yiming from copying his own back road?

For Feng Jianxiong, this is indeed a tricky question.

At least that's what Jack Ma thought.

In the field of artificial intelligence, at least before 2015, because events such as "Google's alpha dog defeated Lee Sedol and defeated the real world champion Ke Jie" have not happened, so the capital community is still relatively rational in throwing money at "deep learning" artificial intelligence.

The price of talents in this field is at least several times, or even dozens of times, lower than later.

The capital tuyere has not yet arrived.

After 2015, anyone who has a Ph.D. degree in the name of "deep learning" (the so-called 'famous school' here refers specifically to Toronto, MIT, Princeton, Stanford, UC Berkeley and other famous schools, Tsinghua University or anything else), has a little project experience or two years (even if it is a project with a supervisor when studying for a doctorate)

Then, Internet giants will basically pay at least 500,000 yuan or more for a monthly salary extension. (If it is at the Tsinghua level, more than 300,000 is fine.) Before 2015, such novices could not get 100,000, which was less than 3 times higher than that of code farmers of the same age/seniority. ）

As for those who have followed Zuckerberg for a long time to do face recognition projects, or followed Google to do intelligent picture engines, and can be the technical service leader of the project on their own, the market after 2015 is millions to tens of millions of dollars in annual salary/dividends/options, these are not dreams.

Who made the artificial intelligence of later generations so popular that it caused panic in the capital circle of all mankind?

It seems that whoever can't catch this bus will be reduced to the bottom of society in the future. All venture capital funds are showing the courage to run away or gamble on their lives, and the value of talents is, of course, fanatical and crazy.

(PS: I've personally met a lot of people who are richer and more educated than me.) When talking about artificial intelligence, they lamented whether their son's generation can survive in the future. Those are people who live in villas worth at least tens of millions, and they are all beginning to worry that their son's generation will not be able to eat. ）

Some people may think: Since artificial intelligence is not so valued and talents are not so valuable, wouldn't it be easier for Feng Jianxiong, who was willing to invest earlier, to spend money and produce results?

But it's not that simple.

Because it is not popular, there are few talents. The scale of cultivation is also small. PROFESSOR JEFF HINTON, THE FOUNDER OF THE "DEEP LEARNING" SYSTEM, ONLY PERFECTED HIS THEORY IN 2006, AND THE FIRST BATCH OF DOCTORAL STUDENTS HE LED ONLY GRADUATED IN 2009, AND ALMOST ALL OF THEM ENTERED GOOGLE AND FACEBOOK.

For professors in other research directions, the progress of talent training will only be slower. So nowadays, there are only a maximum of 3 classes of fresh graduates on the planet who are engaged in this stuff to choose from.

Feng Jianxiong can't recruit anyone with money now. And as long as he wants to raise the price, then those big companies will find a way to stay, or come up with options, or bind people's hearts with future development visions.

Although he has been quietly digging people in the corners this year, in fact, the ones he has dug are all left over from Zhang Yiming's subordinates.

If Feng Jianxiong's action is bigger, he will really be ruthless, so that Zhang Yiming will not even be able to maintain his core business research and development. That Zhang Yiming is not a vegetarian, so it is a big deal to expand financing in advance and reduce the shareholding ratio of the founder team.

In that case, the arms race of the entire industry will escalate ahead of schedule.

This is not what Feng Jianxiong wants to see.

Therefore, he can only dig people moderately and be satisfied with leftovers.

……

People who rely on leftovers to prop up the field, if they say that they are not afraid of technical experts copying their own back, the audience must not be easy to believe.

There has to be a reason to prove that "for some reason, the tech gurus can't replicate my success".

Feng Jianxiong Zhizhu said: "Jack, I believe you should know that Zhang Yiming's "Today's Headlines" is just doing news and information-based content distribution. Videos, even if it's just a short video, he doesn't dabble in anything. With his current funding and expansion rate, he won't be able to expand into video for at least three years. ”

Faced with this statement, Jackma did not react, but just asked Cai Chongxin to respond: "But Xiao Feng, if you want to support your high valuation, you have to prove that 'even if he really becomes big in three years, he still can't cross over' - we are investing money for your future prospects, not for you in the past three years." ”

Feng Jianxiong corrected: "Three years later, he can't turn around if he wants to--artificial intelligence is also divided into many subdivisions, even if it is limited to the one-third of an acre of intelligent push, there are many technical routes that he doesn't want to do."

"Today's Headlines" and Google Push focus on semantic recognition of text content. By letting the algorithm read an article, grab keywords, roughly summarize the core idea, determine what aspects are involved in the consultation, and then push it to readers who may be interested in this field, so as to save the consultation screening time on the reader's mobile phone screen and strengthen the efficiency of advertising/recommendation positions.

And this kind of technology can be said to be useless for the identification of 'video content'. At present, for at least four or five years, no company in the world can judge the tone of a video and the areas involved by 'reading the video content directly' - not even Google.

And now I cut off some of Zhang Yiming's talents and build my own algorithm team, and in the future, I can definitely achieve 'only slightly inferior to Google in this segment, but definitely better than other domestic peers'. The first-mover advantage of artificial intelligence invested by Zhang Yiming in 'reading text content' will not be inherited in the video field in the future. Will go back to the same running line. So, you should admit that I am 'currently the first in the country' in this field and value on that basis. ”

His words are more complicated, but they can be summed up in one meaning: Zhang Yiming is currently stronger than me in artificial intelligence, but his strong direction is different from mine. It's not the same track, it's two tracks that don't interfere with each other.

Jackma groaned.

After all, in 2012, there were not many domestic bigwigs who could deeply understand the route of artificial intelligence technology.

Not to mention that Jack Ma is a marketing talent, even if he is a code farmer like Prednima, he may not be able to understand it.

It is estimated that Boss Li of his mother's family can understand it. (This is limited to discussing the boss, not the technical leader)

Feng Jianxiong said that this was not a track, and Jackma couldn't refute it immediately, so he could only pull twice on his mobile phone and call a technical person who followed him to New York. (Not CTO)

Within two minutes, a bald man in a suit about 50 years old walked into the lounge and politely greeted Jackma and Tsai Chongxin.

Jack Ma was also polite, and relayed Feng Jianxiong's point just now: "Lao Zeng, Xiao Feng just said that Zhang Yiming's technology cannot be used for video recognition and push ......"

After paraphrasing, Jackma asked the man surnamed Zeng to mention his opinion.

Mr. Zeng seemed quite cautious, and did not draw a conclusion immediately, but personally questioned Feng Jianxiong and asked Feng Jianxiong how he came to this conclusion.

After much thought, he finally said to Jack Ma firmly: "Jack, if it is for this reason, I don't think he is right. ”

Jackma's eyes lit up.

You can lower the price!

Feng Jianxiong was not in a hurry, and said with a smile: "You have to talk about evidence." ”

Of course, Mr. Zeng didn't dare to talk nonsense in front of the big boss, and hurriedly laid out the facts and evidence: "As far as I know, although Zhang Yiming is not involved in the push of video content at present, there are other similar push platforms in China that are involved in videos - at least short videos."

As far as I know, there is a company called 'Fast Hoof Technology', which was established last year, the boss is called Su Fat, and their APP is called 'Fast Hoof', which has been developing for almost two years and has tens of millions of users. This APP is for people to share 17-second short videos, shoot and upload them by themselves, and then the platform algorithm pushes these videos to people who may be interested.

There is no doubt that Su Pang's technical route is similar to Zhang Yiming's, and he is involved in short videos - this shows that artificial intelligence push can be combined with the video industry at present, why do you still say that you are firmly in the first place in China?"

However, after Mr. Zeng finished speaking, Feng Jianxiong didn't have time to answer. However, Jackma and Cai Chongxin were curious in unison:

"What kind of company is this? Why haven't I heard of it? There are tens of millions of users' products, and there are products that I haven't heard of?"

For a moment, Jackma actually had the illusion that he was too incompetent, and that he was so obsessed with the dynamics of the industry.

Mr. Zeng hurriedly teased to the boss: "Jack, it's okay if you don't know, the main thing is to blame this company's main market, so the reader basically doesn't use it." At present, almost all of the 20 million users are migrant workers and couriers, as well as left-behind women with children in rural areas.

The most trendy user group is probably the washing, cutting and blowing at the junction of urban and rural areas, accounting for about 10 tens of thousands, as well as nightclub girls and factory girls who have just come out to mix, adding up to millions, which is the main force of the content provided by this APP. ”

It's very straightforward, and it makes things clear, but it's not very good.

Of course, Jackma is a person who wants to be personable, and after he absorbs the content nutrition, he has to pretend to be a good person: "Alas, Lao Zeng, you can't say that." There is no distinction between high and low occupations, but the division of labor is different. There are 20 million people using something, I don't know, it's my job is not in place, I should introspect. ”

However, looking at Jackma's attitude, he is already very satisfied: with the argument provided by Mr. Zeng, Feng Jianxiong's position that "the current domestic artificial intelligence push algorithm can't do video well" is untenable.

Facts speak louder than words!

You Feng Jianxiong said how different the technical route is and how it cannot be inherited or transplanted......

Even if it is said in a fanciful way, but the facts are in front of you, that is, there are peers who have done it, and you are still a hard fart?

No matter how hard you are and your theoretical skills, you can't bear the objective facts!

Jack Ma seems to have seen the achievement of "bargaining success" and is already beckoning to him.

In fact, he is not reluctant to give up these hundreds of millions of dollars.

It's just that he pretended to be the king in his life, and he was born as a teacher. In the past 20 years, he has never been the only one who has been selling his mouth in front of others to win profits.

But as soon as he met Feng Jianxiong, he could only be persuaded by Feng Jianxiong and spray to win.

Everyone has also had six years of cooperation and friendship, and in Jackma's memory, he never won Feng Jianxiong.

It's so unpleasant!

This time it must be turned over!

"Heh, hehe, hahahahaha......" Feng Jianxiong didn't have any awareness of the imminent disaster, and continued to taunt. His laughter was so rampant and confident that it made people's hearts bristle,

"Mr. Zeng, Mr. Zeng, you care more about the forefront of the industry than Lao Ma and Lao Cai, this is your point. But I don't know if you've ever heard a saying: what you get on paper is shallow, and you never know that you have to do it.

You just listened to the name of 'fast hooves', and then listened to their own propaganda of the technical route, and they all believed it? You want facts to speak louder than words, yes, we are now under our own APP, I am on my mobile phone up and down today's headlines, you on your mobile phone up and down fast hooves.

Let's follow the same tonality to 'tune' the algorithm to push our personalized content. The same number of 'likes' and 'dislikes/dislikes' at the same time. After the time is up, whether these two algorithms have been tuned by 'big data'. I'm going to show you if it's true?

"Xiao Feng, you ......" Zeng was always a little unhappy, but he also knew that what he said just now was a bit arbitrary and offended the other party.

Since you want to use facts to make an appointment, beeping is useless, so let's do it.

Mr. Zeng gritted his teeth and responded: "Okay, let's train according to the same standard - this way, every content or video, like it or not, will be judged according to Mr. Ma's preferences." That's the fairest way to go. ”

The two said they would do it, downloaded the APP on the spot, and then began to adjust the algorithm preferences.

Jackma took turns to glance at Lao Zeng's mobile phone, and then at Feng Jianxiong's mobile phone, and said "like/favorite" or "dislike/disinterested" in turn.

The function of "stepping" in "Today's Headlines" is relatively fine, and the reasons for disliking it could have been further subdivided into many categories such as "repetitive content/vulgar/disgusting discomfort/blocking the author".

However, because there are no such functions in the fast hoof, I was afraid that Feng Jianxiong would take advantage, so I asked Feng Jianxiong not to give specific subdivision reasons, so he could only choose "I don't like it" in general.

Feng Jianxiong was not frightened at all and gladly accepted this kind of competition rules.

Half an hour later.

Jack Ma took the two mobile phones and re-swiped the APP.

He thought for a while, and said very measuredly: "Lao Zeng, it seems that you lost - in today's headlines, the content I am interested in has been adjusted for half an hour." Now I have swiped the first 15 articles, and there are already 11 of them that I am interested in clicking in. Fast hooves here, brush out the first 15, I still only have 3 barely read. ”

Lao Zeng was still unconvinced, and finally defended: "Mr. Ma, you can't blame me, it's just a bit of style in the fast hoof, and I can't make content out of nothing if I desperately adjust the algorithm." ”

"Lao Zeng, don't quibble—I've counted historical data. When we first started to tune the algorithm just now, when we swiped out 15 new pieces of content several times, Mr. Ma already had 3 or even 4 of them that he was interested in.

So, your quibble doesn't hold up – the content is too tonal and can't be used to explain 'why the algorithm hasn't improved in the slightest'. If at the beginning, Ma always thought that there was 1 or 2 out of 15, and now it has risen to 3 or 4. Then you can still explain.

But now? It's not a question of absolute numbers, it's a question of whether there is progress - at least in my opinion, this kind of propaganda in the name of artificial intelligence is just a swagger, and its algorithm is not well tuned at all. ”

Lao Zeng was speechless.

It's like a student's poor test scores, which doesn't prove that the teacher is poor - maybe the student is just a bad guy and can't get on the wall?

However, if a person who claims to be a famous teacher has been trained for a long time and his grades have not improved, there is a problem. Last year's test was 20 points, this year's midterm is still 20 points, and the final semester is still 20 points, so what is the use of the teacher?

Jackma felt that he couldn't understand this question, and he was really uncomfortable, so he was ashamed to ask: "Xiao Feng, you explain why Kuaiho's algorithm can't do 'intelligent push video'? As long as you say that you are ugly and can prove that your domestic counterparts in this field are weaker than you, I will admit your valuation today! I will take an extra 500 billion US dollars and listen to you take a lesson on the frontiers of artificial intelligence!"

Zhou Tianyin, who was always in a daze listening posture on the side, couldn't help but get nervous when he heard the words.

That's a multi-billion-dollar price difference.

She involuntarily pinched Feng Jianxiong's arm.

Feng Jianxiong gently wiped away Zhou Tianyin's hand and signaled her not to worry.

Don't stop my brother, my brother is going to pretend to be forced.

"This starts with the technical route of Su Fat......"