When Zhang Yiming encounters a "non-compete agreement"

Some things are not worth weighing, but once they are, they are immeasurable.

In the past few years, internet workers who have encountered mass non-compete agreements should have a more personal understanding of this sentence. A large portion of departing employees who signed non-compete agreements never anticipated that one day this piece of paper would truly make them pay a huge price. After all, they still have some value, and the company's high-handed approach is incalculable.

The legal basis for non-compete agreements is Article 24 of the Labor Contract Law, which states that non-compete restrictions only apply to "senior management, senior technical personnel, and other personnel with confidentiality obligations of the employer." However, the reality is that more and more grassroots employees are being included in the scope of non-compete restrictions.

An article in "LatePost" earlier this month cited a case where a cross-border e-commerce company employed about 50,000 outsourced laborers, and the new contracts they signed with the company in February added non-compete clauses. In the usual understanding, outsourced labor and the "two highs and one secret" roles in regulations are obviously quite distant.

However, in recent days, a real senior management and senior technical personnel have encountered a similar situation.

On April 19, "The New York Times" published an article stating that a Pennsylvania court mistakenly made sealed court documents related to ByteDance public. It mentioned that two contractors accused Hina International of bringing some cutting-edge search technology to ByteDance without providing reasonable compensation.

Zhang Yiming previously worked at Kuxun and met Wang Qiong from Hina Asia through this connection. Later, Wang Qiong invited Zhang Yiming to start a vertical real estate search website, Jiufang. The entrepreneurial experience of Jiufang clearly had a direct inspiration for Zhang Yiming to start ByteDance, and the focus of the above-mentioned lawsuit is whether there was improper transfer of technology during this process.

Jiufang's lawsuit

Jiufang was founded in September 2009 and was once the largest real estate search website in China. Zhang Yiming's idea for Jiufang at the time was to use different products to serve different user needs, so Jiufang launched a bunch of apps, including "Real Estate Information," "Renting on the Go," and "Buying on the Go," among six other products.

Later, when ByteDance transformed into an "APP factory," some people traced this practice back to the era of 99fang, but fundamentally, these are two completely different approaches.

Like Baidu, although 99fang is a vertical search engine for the real estate industry, it does not produce content itself, but relies on web crawlers to collect public information from the internet, such as posts on 58.com or Ganji.com. This approach effectively solved the cold start problem of the website and hit the pain points of users at that time. Because at that time, the mobile internet was just emerging, and the industry had not yet figured out how to do real estate brokerage products.

However, after the initial start-up phase, 99fang did not switch to a platform operation model and still did not introduce strong operation and review mechanisms. Just like today, the authenticity of various rental information on major platforms is difficult to discern, and the problem of fake listings would have been more prominent at that time.

But it may not be that Zhang Yiming didn't want to do it, because this was not just a problem for 99fang. Platforms like Ai Wu Ji Wu also once had a strong presence, but eventually disappeared. In the words of Zuo Hui, "Consumer internet goes horizontal before vertical, while industrial internet goes vertical before horizontal." Without the accumulation of deep industry cultivation before, relying solely on the internet cannot change the fate of Lianjia.

The 99fang project cannot be considered very successful. In the case of a matrix-style product layout, it only had 100,000 daily active users by 2011. Zhang Yiming sees entrepreneurship as gambling. Since success is a low-probability event, he wants to increase the frequency to increase the probability of success. In this situation, in 2012, he found a new CEO for 99fang and founded ByteDance with a few people.

Among the pile of APPs in 99fang, the one that had a profound impact on all of ByteDance's later products was "Real Estate Information." "Real Estate Information" can basically be seen as the early product prototype of today's Toutiao, except that the content focuses on the real estate vertical field.

The first product director of Today's Headlines, Huang He, described the app as "collecting and distributing all real estate information, which is especially popular. First, aggregate the information, and then make recommendations. You can even understand it as the real estate channel of Today's Headlines."

The operation mode of real estate information has become the framework for all products of ByteDance later on. Whether it's the initial "funny pictures" and "content jokes," or the later "Today's Headlines," they are all built on the actions of content crawling, cleaning, aggregation, and recommendation.

In the beginning, Jiufang once sparked controversy in the outside world because it crawled real estate sources and content information from other websites. Later, copyright issues also became the biggest crisis that Today's Headlines faced in the early days.

These origins all prove the close connection between Jiufang and ByteDance, but it is not tenable to determine that ByteDance inappropriately obtained certain intellectual property from Jiufang.

The article in The New York Times did not mention the specific demands of the plaintiff, and the relevant court documents have been resealed and cannot be accessed. However, by sorting out Zhang Yiming's entrepreneurial experience and ideas, we can still make a rough judgment on this war of words.

Does Zhang Yiming need "technical assistance"?

The article in The New York Times revealed that the specific allegations made by the plaintiff were related to "cutting-edge search technology." Whether it was the earlier Kuxun or Jiufang, search technology was the underlying support of the product.

Zhang Yiming once summarized three characteristics of Kuxun's search, including city-focused local search, high real-time instant search, and accurate response to user needs for vertical search. At that time, Zhang Yiming was the chairman of the technical committee at Kuxun.

In other words, at least during the period of Kuxun, Zhang Yiming's technical accumulation in search engines was already sufficient to support the operation of a large website.

Jiufang was initially split from Kuxun's real estate channel. From this perspective, there is obviously no reason to believe that Zhang Yiming, who later went to MSRA, then worked as a technical partner with Wang Xing for a period of time at Fanfou and Hainei.com, needed technical assistance from investors in the United States when he was doing Jiufang.

Of course, considering that Haina had invested in Kuxun before, it cannot be ruled out that "technical assistance" was provided during the Kuxun period. However, if this is the case, it is quite far-fetched to associate this with ByteDance, which was established in 2012.

"Shendu" tends to believe that in Zhang Yiming's career from Kuxun, Haina may have introduced some technical resources from the outside from the perspective of post-investment services, and relevant compensation agreements had already been negotiated at that time. However, when ByteDance soared to become a giant worth tens of billions of dollars with TikTok and Douyin, the former technical contractors felt that they deserved a share, as the cake was too big and even a small piece of the return would be substantial.

In fact, in the field of coding, it is unrealistic to monopolize technology that surpasses others, whether from the overall industry atmosphere or practical operability. The decisive factor has never been technology, but rather the entrepreneur's judgment of trends and the ability to quickly iterate products.

The internet does belong to the category of high-tech industries, as can be seen from the generally good educational backgrounds of internet company founders. When the user base of a product expands rapidly, it often requires the reconstruction of the original rough software. With the increase in concurrent users, a CTO with insufficient expertise will inevitably struggle.

However, it is worth noting that this is also the most "knowledge-equal" industry. In this field, the most cutting-edge technological advancements, whether it's early operating systems like Unix or recent large language models, have all been made public, despite being know-how that should have been kept secret in traditional industries. Therefore, it seems that in the past twenty years, no internet company has purely gone out of business due to technological constraints. The best practices of the industry are there for all to see, and even if you don't know how to do it yourself, can't you hire someone who does?

On the other hand, from a technological perspective, what really laid the foundation for the success of Toutiao and Douyin today was the recommendation engine, not the search engine. Although both involve complex data processing processes at the implementation level, the goals they seek to achieve are completely different.

Traditionally, search engines require fast and accurate responses because they are triggered by explicit user queries, and there is not much room for personalization. Recommendation engines, on the other hand, mainly predict content that users may be interested in based on their behavior and preferences, offering greater freedom and personalization. In general, one is a passive response, the other is active recommendation; one is user-driven, the other is system-driven.

Zhang Yiming recognized the value of the recommendation engine very early on. During his time at Kuxun, he wrote a program that automatically searched for train ticket information and sent him a text message if the conditions were met. When he later recalled this experience, he said that the paradigm shift from people finding information to information finding people was his earliest thinking and practice regarding recommendation engines. Later, with the "Real Estate Information" product at 99fang, he had the opportunity to practice and strengthen the information distribution approach based on recommendation engines.

But in terms of the maturity of recommendation engine technology, Zhang Yiming was still in a "learning while doing" state until the launch of Toutiao.

At the end of 2012, in the office on the sixth floor of Jin Qiu Jiayuan, Zhang Yiming convened a meeting with the team, with the core topic being "to build an information platform, it is necessary to do a good job of personalized recommendation engines, should we start this now?"

When he was encouraging the team, he also said a sentence: "We don't know about recommendations, but we can learn." Later, when he found out that Hulu's Xiang Liang was writing a book called "Practical Recommendation Systems," Zhang Yiming even tried to get an electronic version from him to study, although he was refused on the grounds that the book had not been published. These details illustrate two points: opportunities are always left to those who are prepared, but there are no completely prepared people; and when ByteDance started its business, Zhang Yiming and others were still somewhat rushing into it.

Conclusion

Zhang Yiming once said a sentence, "Your understanding of things is your greatest competitive advantage." For Toutiao, Douyin, and TikTok to have come this far, technology is an important factor, but the shift in understanding from "people finding information" to "information finding people" is the decisive factor. From this perspective, when the content carrier of the internet switched from text and images to videos due to the upgrade of communication infrastructure, it seems that ByteDance's incubation of Douyin from Toutiao was inevitable.

Therefore, as mentioned in "Shendu," the purpose of the plaintiff initiating this lawsuit seems more like a naked and far-fetched "grab for the cake." However, from another perspective, are the increasingly strict non-compete agreements of major internet companies today also similarly naked and far-fetched? Haina International has ample resources to deal with litigation, while workers facing similar situations are in an absolutely weak position in the game against big companies.

This is worth reflecting on.

*The pictures in the article are from the internet.