The next day, while Lin Feng silently worked on a recommendation algorithm and waited for Great Xia Armaments Industry and overseas military contractors to come negotiate with him,
by noon, many Bilibili uploaders who resented and envied Lin Feng had already drawn their blades. The leader was none other than the famous Director Ao.
Director Ao was truly one of Bilibili's old guard. Although his main video platform was Youku, after all, Youku had more users.
Still, he had always kept an eye on Bilibili, and whenever he uploaded a video, he would post it on Bilibili as well. It could be said that he had contributed enormously to Bilibili.
In his previous life, when Bilibili users evaluated the five uploaders who had made tremendous contributions to Bilibili's development, Director Ao had been one of them. This showed just how much he had contributed.
Back to Director Ao. Earlier, when he saw a newcomer gain 400,000 followers in just one day, rising even faster than he had in his prime, he became deeply jealous.
However, as a well-known public figure, he had not been in a position to speak out when others were making parody videos of Lin Feng and videos criticizing him.
Now, as he watched Lin Feng develop better and better, even about to obtain 30% of Bilibili's shares through a wager agreement and a recommendation algorithm, Director Ao felt increasingly unbalanced.
Comparing people could drive one mad. I had started earlier than you—why did you rise so quickly and gain so much attention and so many followers?
Driven by jealousy and resentment, Director Ao began staying up that night to gather all kinds of information on recommendation algorithms. The more he researched, the wider the smile on his face became.
After working for most of the night and the entire morning, a carefully produced video was officially released on video apps such as Youku, Tudou, Tencent, AcFun, and Bilibili.
"Hello, everyone. I'm Director Ao. We meet again.
A lot has happened on Bilibili these past few days. First, the website and app were redesigned, with graphic ads added, and then a livestreaming platform was launched...
What surprised me most was that Lin Feng actually signed a wager agreement with Bilibili's shareholders. I also watched last night's livestream.
During it, I noticed the recommendation algorithm Lin Feng claimed was worth 20% of Bilibili's shares. I imagine everyone must be very curious about this matter.
So, after staying up all night gathering information and organizing it the following morning, Director Ao has lived up to expectations and made this video."
By this point in the video, Bilibili was filled with bullet comments such as, "Director Ao has worked so hard," "Director Ao is so dedicated," and "Director Ao, take care of your health."
It was clear that Director Ao still had many fans on Bilibili. They loved his videos and, by extension, cared deeply for and admired him.
"All right, let's get to the point.
This video will mainly introduce recommendation algorithms, then estimate whether the recommendation algorithm Lin Feng mentioned is truly worth 20% of Bilibili's shares."
Upon hearing this, Director Ao's fans grew increasingly excited. They were extremely curious about the mysterious recommendation algorithm.
Meanwhile, Director Ao in the video began steadily explaining recommendation algorithms:
"First, recommendation algorithms originated at the Palo Alto Research Center in 1992. They created a system based on collaborative filtering algorithms and used it for spam filtering.
Then, in 1994, the University of Massachusetts and the University of Nice introduced a news recommendation system based on collaborative filtering algorithms, which could recommend news according to users' ratings.
However, recommendation algorithms were not truly applied to the internet until 2003. They were used by e-commerce platforms, which employed collaborative filtering algorithms to recommend similar products.
Then, in 2006, Douban and Netflix of the American Empire introduced the world-changing matrix factorization method. Both users and products were assigned corresponding latent vectors, giving it strong generalization capabilities.
From then on, e-commerce product recommendations became more targeted, and the probability of users purchasing successfully recommended products rose greatly.
Finally, in 2010, Osaka University in Japan proposed the FM model, or factorization machine, a machine-learning model. Recommendation algorithms officially entered the era of machine learning.
It was particularly suited to handling feature interaction problems in sparse datasets. From then on, product and advertising pushes became more precise and efficient. After two years of development following last year, 2012, it officially came to dominate the recommendation system field.
One thing worth mentioning is that, with the release of the AlexNet convolutional neural network last year, many people also set their sights on neural networks.
They wanted to rely on neural networks to develop new recommendation algorithm models and open up new avenues of technological development.
But let's not even talk about recommendation algorithm models based on neural networks. That is an entirely new technical path, and not a single finished product has yet been released.
Even the now quite mature FM model can achieve what Lin Feng described: precisely delivering videos users might be interested in to consumers.
So recommendation algorithms are not some magical thing. They have existed for more than twenty years and have become quite mature by now.
I suspect that Lin Feng developed Bilibili's recommendation system based on the FM model, ultimately achieving the goal of recommending videos.
This kind of technology is not particularly exaggerated. Honestly, there are at least hundreds, if not a thousand, people in China capable of researching this kind of recommendation system technology, and there would only be more overseas.
Lin Feng claims his recommendation algorithm is worth 20% of Bilibili's shares. I think that is somewhat overconfident. Sometimes confidence is a good thing, but excessive confidence is arrogance.
Of course, we can understand why Lin Feng is so confident. After all, Lin Feng himself said last night that his algorithm research had not officially begun yet.
It's like how all of us, while we were in school, always said we would definitely earn an annual salary of a hundred thousand in the future, buy a car, and buy a house. Then, after entering society and starting work, we realized just how naive we had been.
Lin Feng's algorithm has not officially been researched yet, so it is normal that he does not understand how algorithms work. We should understand Lin Feng."
Director Ao kept saying good things about Lin Feng, and there was not a single curse word in the entire video. Yet every viewer knew that Director Ao was secretly mocking Lin Feng for being too arrogant.
Naturally, this caused the bullet comments and comment section to split into two opposing camps, with both sides possessing massive fanbases.
Lin Feng had made a name for himself through inventions such as the Handwriting Robot and Coin-style Electromagnetic Rifle, as well as two lavishly produced documentaries and I Love Invention. He had successfully gained millions of followers.
Although Director Ao's experiences over the past few years had not been as spectacular as Lin Feng's, he had also accumulated millions of followers over the years since he began making videos in 2009.
The two had roughly the same number of followers, but Director Ao had more loyal fans. After hundreds of videos, people had long become his die-hard supporters.
Thus, Lin Feng's fans were quickly defeated and simply could not compete with Director Ao's fans. How could one mouth fight against several?
This afternoon, I received notice of a second round of recommendations. Overjoyed, this author has decided to release an extra chapter as thanks for everyone's support.
Also, please send recommendation votes and monthly votes.
Before you continue