ByteDance is one of the greatest companies in the history of the internet. Its core, the recommendation algorithm, has essentially led the entire mobile era of information distribution. Today, it is almost impossible to find any major information application that does not use algorithms, and also hard to find one that does not set algorithm-driven feed as the default option.
It can be said that almost all the upgrades in information distribution efficiency over the past fifteen years revolve around recommendation algorithms.
Of course, a core reason for it to thrive is the explosion of mobile technology. This allowed an electronic device to become an extension of a person, or even become a part of that person. It also made it possible to collect more and broader user data, which serves as the fuel for more accurate algorithms.
In the lifecycle of consumable information, distribution is a largely overlooked part that requires effort. In fact, good distribution requires as much effort as good content creation to achieve good consumption results. This is one of the foundations of the success of recommendation algorithms: they produce a highly effective distribution outcome (information closely related to each user) with minimal human effort.
Before algorithms, if a user wanted a good distribution result, they typically had to rely on a few professional editorial teams to select the content that best suited their preferences, hoping that the editors would consistently make good judgments. We can temporarily call this Editor-Based Distribution.
Alternatively, users could manually search for and filter out quality information and content themselves, becoming their own editors. We can call this Self-Based Distribution.
In the diagrams above, the thick red line represents the distribution process requiring effort, and the circles connected to the lower end of the red line is the target of this effort.
Among these, editor-based distribution occupies the majority in terms of user coverage and engagement. The reason is simple: the reward for the self-based model is merely "information that aligns with one's preferences and efficiency," while the effort required is immense.
With such premise, the invention of recommendation algorithms was a devastating blow to both of the above models. All human participants in the system could achieve fairly good information distribution results with minimal effort.
As a result, algorithm-based distribution now controls almost the entire internet. Meanwhile, this distribution model has brought numerous downsides—reinforcing biases, information echo chambers, sensational content, excessive pursuit of traffic, and more. These issues have long been topics of discussion among practitioners and users alike, leading to major complaints about algorithms.
Collectively, users will always seek to achieve the best results (products, experiences, etc.) with the least effort and cost (financial, cognitive, physical actions, etc.). Therefore, even though people have many complaints against algorithms, they continue to dominate the world. Only a small fraction of users still rely on editors or their own efforts to process information.
Here, we focus on the latter: those who still maintain the habit of actively searching for and organizing information. These users still utilize methods like searching, subscribing, unsubscribing, and filtering to ensure the quality and accuracy of their information intake. Although their numbers are small compared to the audience of algorithms, they are still significant in absolute terms (e.g., the global user count of a single RSS client exceeds 20 million).
This leads me to ponder: If distribution itself is a form of effort, and there has always been a substantial group of users actively working to enhance their information experience, could we view their efforts as a work that others can consume?
If content creation is a kind of effort that can be monetized, could information distribution also be a form of effort that can be monetized? This way, many efforts that couldn't be monetized before could be incentivized, making the entire distribution system unprecedentedly decentralized. Here, decentralization refers to social rather than technological aspects (for more on technological decentralization, visit https://rss3.io/).
As a minimal mechanism, if someone has a specific subscription list and keeps maintaining it to best serve themselves and a small group of similar people, wouldn’t they be willing to share it in exchange for some social recognition? If regular users can easily consume these curated, real-time updated information feed, would this not provide more value than the existing distribution systems?
For ordinary users, this offers a way to gain a good distribution experience without putting in effort, just like with editors or algorithms. At the same time, it avoids the limited options brought by editors and the high echo downsides created by algorithms.
For users who are already organizing information, their efforts can not only enhance their information consumption efficiency, but also serve as a social capital, a financial capital, or even a currency to gain benefits and profits.
An effort-maker and their users can form an organic connection. Rather than just letting algorithms observe your behavior to adjust parameters, you can interact with real people to provide feedbacks or even become an effort-maker yourself, helping more people and gaining rewards in the process.
Of course, in the actual product, we still need to address some cold start problems. For instance, if information distribution effort is a commodity, then this market must have enough "sellers" and "buyers." In this two-sided market, “sellers” are the user experience of “buyers”, and how to get them to join, understand this system, and better serve themselves and others is crucial.
If Pinduoduo, to some extent, changes e-commerce through capitalizing socializations, can we propose a possibility to change information consumption through capitalizing distributions?
Could we also introduce decentralized technologies and, along with this decentralized distribution system, attempt to simultaneously solve distribution's capitalization, monetization, and further enhance the healthiness, efficiency, and sustainability of the Web?
Of course, building such a system involves two other core factors: having a large enough open information layer available in an era of information monopoly and offering a truly great end-user experience.
If algorithms are centralized efficiency brought about by machine evolution, then this new capitalized distribution model might be a possibility for humans, as a vibrant society, to fight against it.