Those who are changing the world with recommendation engines

There are always people who bring tears to our eyes. 1 A Bei and Wang Shoukun want to use the recommendation engine to change the world. I don’t know whether young people are still playing Douban. They should not know that Douban is the first company to put forward the slogan “recommendation engine changes the world”. However, they may still have an impression of China’s number one literary and artistic youth + idealist “Douban Abei”. Abei (Yang Bo) joined IBM, which invented the computer, as a consultant scientist in 1998. Two years later, he joined a Python mailing group, and discovered activist Hong Qiangning. At the end of 2005, after Abe founded Douban, he planned to bring Hong Qiangning into the company, so he gave him an interview question to solve the bug. After Hong Qiangning completed it in Python, he became the No. 2 employee of Douban. Hong Qiangning recalled the scene of the meeting that day: “Haha, Abe used Java originally, but when he was doing Douban, he said that he hated Java, so we used Python from the beginning.” He later took the initiative to cut his salary to join Douban , as the chief architect and an outstanding Python evangelist in the early days of China. Before getting the financing, the entire Douban had a boss from Abe and a regular employee of Hong Qiangning (Brant, who was in charge of community management, was still a part-time employee at the time). The two met in a cafe once a week, chatted until midnight, and then went home. Work until June 2006 when they moved to an office next to Beijing’s 798 Art District. It has been widely rumored on the Internet that Douban was named after the company’s location, “Douban Hutong”, but Hong Qiangning said that he just passed by there and did not stay there, only Abei lived there for a short time. From the first day of Douban’s establishment, the problem that Abei wanted to solve was too narrow information. He clearly realized that recommendation is an effective solution. Like Zhang Yiming later, Abei wrote the first version of the recommendation engine by himself, which is also the first recommendation engine in the Chinese Internet world. From the point of view of recommendation effect alone, this recommendation engine is actually quite powerful. At that time, the main users of Douban were hard-core literary youths who loved to read and create, and they desperately needed a tool to explore and discover new books. Abe helped them solve this need, so Douban became popular in the blogosphere and gained a lot of traffic. However, from the perspective of operating efficiency, Abe, a pure Python recommendation engine, is also slag. With the continuous influx of users, it often takes a whole day to calculate the results. However, Abei is not as worried about the lack of technology as Zhang Yiming, because he has Wang Shoukun, an old friend who has been helping him for many years. In fact, Wang Shoukun was holding two offers at that time, one from Amazon, the world’s richest e-commerce company, and the other from Douban, the world’s most sentimental company. These two offers need him to do. It’s all the same – that’s the recommendation algorithm. This is not a choice that needs to be considered, because Wang Shoukun stood by Abei without hesitation, just because Douban looks more interesting. Teacher Wang Shoukun recalled his work when he first joined: “After I joined the No. 4 employee, I only did one thing in the first two months, which was to change the pure Python version of Abei to the pure C version, and then the recommendation efficiency was probably improved. It looks more than ten times.” According to Wang Shoukun, Abei’s version is still very good, and he thinks that what he is doing is mainly to optimize performance. Also in 2005, Xu Yirong, who was also from IBM and also found it interesting to recommend, founded Shrimp. He gathered 30 engineers and grabbed 1.3 billion articles from more than 500 blogs and news channels. Education and catering were all included in his shrimp catching reader, and he even began to comment on his brother Li Yanhong, who was starting a business at the time, “nothing special.” If we use the term of the later mobile Internet era, what Xu Yirong wanted to do was a recommendation engine focusing on information distribution, but he never really built a recommendation tool, and he still used the traditional keyword rule matching method from the beginning to the end, mainly Rely on a good product experience to acquire customers. In addition, Xu Yirong’s idea was too big, and he did it too early. He caught shrimp and finally caught blind, so he could only sell it to Douban. Although catching shrimp did not bring much technology to Douban, it also contributed a lot of content re