Google's main source of income is advertising, and the main source of advertising revenue is search recommendations. The core of search recommendations is the ranking of relevant web pages.

The reason for using competitive ranking is that users' attention is limited, just like wealth. The top few pages will receive over 95% of the user's attention.

The number of searches by ordinary users and the number of pages read each day are limited. The core issue of recommended advertising algorithms is how to push the most effective ads within a limited number of pages.

All advertisers know that at least 70-80% of their advertising money is wasted.

The problem is that no one knows which 20-30% actually works.

Not a penny less should be paid.

For Google, how to better plan and allocate these display interfaces is a way to make a profit without any investment.

For example, if Google can push ads to everyone that are exactly what they need and will click on, it will achieve the theoretical limit of recommended advertising, earning all the money with the fewest displays.

Unfortunately, most companies can't achieve even 1% of this ratio, and even Google is only around 2%.

Meng Fanqi's "mobile optimization strategy" and "Fred algorithm" are designed to further eliminate web pages that are inconvenient to view on mobile phones and various traffic garbage websites.

It's a process of elimination. As long as I exclude more junk content, the remaining content is more likely to be clicked on.

And combined with the language AI model, it's an additive process, providing more relevant results by better understanding the user's search intent.

As a result, the interface seen by users after each search is more relevant, with fewer junk websites, resulting in more ad clicks.

Based on the experience of the original timeline, Meng Fanqi estimated that after implementing these three strategies, the improvement should start at around 20%.

As Hixiaohu listened, he found the details of the plan more and more interesting. Originally planned for just a few minutes, they had been talking for nearly an hour without realizing it.

The two of them didn't find a meeting room, but instead found a sofa near the front desk and chatted.

Many Google employees passing by on their way to work were quite surprised.

Hixiaohu's position was quite high, and not many Google employees had direct contact with him.

A few interns had hardly even seen this big shot.

"Hey! Zhou Huan, Chen Hui, come over here." Hixiaohu caught sight of two passing Google employees and called them over.

Zhou Huan and Chen Hui had just gone to get coffee and were curious to see the big boss chatting with a young student-looking person.

On the Baidu side, due to the immediate promotion of the product, Li Yanhong personally did two promotions, and many people knew Meng Fanqi.

On the Google side, Jeff and Hinton had only introduced Meng Fanqi within the Google Brain, and it didn't involve too many unrelated personnel.

Therefore, Zhou Huan and Chen Hui, who were focused on the advertising part, were not clear about Meng Fanqi's background. Seeing him so young and chatting with the big boss, they thought he was an intern who got in through connections.

"Leader! You can rest assured, we will definitely take care of him!" Hixiaohu had just introduced a name, and Zhou Huan preemptively expressed his willingness to help.

"Alright, you two prepare the local relevant accounts, permissions, and tools. He will mainly be working on our search recommendation department during this time." Zhou Huan and Chen Hui were both experienced engineers in the group, had a sense of proportion in terms of technology, and knew how to lead people.

Hixiaohu felt that although Meng Fanqi's ideas were quite meticulous and excellent, it was still his first time entering the advertising algorithm department to write core content.

Leaving aside other things, it would take quite some time to familiarize himself with all these existing things.

After all, understanding the principles and mastering the code to develop and create on his own was a world of difference.

He could see that the two misunderstood Meng Fanqi's identity, but he didn't hint at anything.

One was because Meng Fanqi did indeed need to familiarize himself with a lot of content like an intern.

The other was because Zhou Huan and Chen Hui, who were always a bit clever, often had their heads in the clouds and often caused some trouble.

The two of them had made quite a few contributions before, so he didn't say much.

"Hmph! This time, let you guys see what real talent looks like." Watching Meng Fanqi walk to the work area with the two, Hixiaohu sneered.

"Little brother, how old are you this year? How far have you read?" On the way to the workstation, Zhou Huan began to inquire.

"I'm in my early twenties, and I'll be a senior next year." Meng Fanqi answered truthfully.

"Good!" Zhou Huan cheered loudly in his heart. He had already made a bet with Chen Hui when he saw Meng Fanqi chatting with the leader privately.

The bet was whether Meng Fanqi was an intern or not, and now Zhou Huan thought he had won hands down.

As they continued walking forward with the two, Zhou Huan put his left hand behind his back, moved four fingers, and signaled Chen Hui to admit defeat.

Chen Hui rolled his eyes and took out his phone to send Zhou Huan a 200 yuan red envelope.

"Have you arranged your internal account?" Zhou Huan saw that Meng Fanqi had logged into the internal Google account shortly after sitting down, and he felt a little uneasy.

"Yes, now I mainly need you to introduce the process of putting the advertising algorithm online to me. I need to obtain all the relevant data and script permissions in this process."

Meng Fanqi probably knew about this testing process. Usually, it would be tested offline on different scales first, then partially launched, and after testing normally, it would be implemented globally.

Backups would be kept at each stage in case of problems that required a rollback at any time.

"I'm super, Google Brain?" When Zhou Huan saw Meng Fanqi's email suffix while assigning permissions, he was taken aback.At this time, although Google Brain is not yet the behemoth team it would become later, it still has some reputation due to the leadership of star engineer Jeff.

Chen Huizui smirked and looked smugly at Zhou Huan, according to the rules, he would ask for money without clarifying the situation first, and this would have to be returned double.

Zhou Huan shouted in frustration, and reluctantly handed over 200, then transferred another 400.

Meng Fanqi glanced sideways, probably understanding what these two guys were up to.

Capitalist companies are not aggressive enough, always spending their energy on trivial matters.

Explore the extended edition on pawread.com.

Meng Fanqi, who received the authorization, was the first to implement the Fred algorithm.

The AI-related part requires some time for preparation and training, and the method for optimizing the mobile end may not show significant results in the short term. The effect will be better later in 2014.

But the Fred algorithm's ability to filter out internet junk websites is quite remarkable.

After the algorithm goes online, nearly 70% of the websites will likely be greatly affected.