18 August 2020

Remains of the Day: “TikTok and the Sorting Hat”

More than any other feed algorithm I can recall, Bytedance’s short video algorithm fulfilled these two requirements. It is a rapid, hyper-efficient matchmaker. Merely by watching some videos, and without having to follow or friend anyone, you can quickly train TikTok on what you like. In the two sided entertainment network that is TikTok, the algorithm acts as a rapid, efficient market maker, connecting videos with the audiences they’re destined to delight. The algorithm allows this to happen without an explicit follower graph.

Just as importantly, by personalizing everyone’s FYP feeds, TikTok helped to keep these distinct subcultures, with their different tastes, separated. One person’s cringe is another person’s pleasure, but figuring out which is which is no small feat.

TikTok doesn’t bump into the negative network effects of using a social graph at scale because it doesn't really have one. It is more of a pure interest graph, one derived from its short video content, and the beauty is its algorithm is so efficient that its interest graph can be assembled without imposing much of a burden on the user at all. It is passive personalization, learning through consumption. Because the videos are so short, the volume of training data a user provides per unit of time is high. Because the videos are entertaining, this training process feels effortless, even enjoyable, for the user.

I like to say that when you gaze into TikTok, TikTok gazes into you.

Eugene Wei

I have read this long article before ‘The TikTok War’, so I appreciated the context about how TikTok evolved over the years from Musical.ly to current owner Bytedance. I also appreciated the distinction of categorizing TikTok as an ‘entertainment network’ as opposed to regular ‘social networks’ because its main use case is not to connect to other people but rather to consume content.

A few years ago, on a visit to Beijing, I caught up with a bunch of former colleagues from Hulu Beijing, and all of them showed me their Douyin feeds. They described the app as frighteningly addictive and the algorithm as eerily perceptive. More than one of them said they had to delete the app off their phone for months at a time because they were losing an hour or two every night just lying in bed watching videos.

But I find other ideas expressed here deeply disturbing, dare I say contrary to a democratic society. In the paragraph below, the author argues that people are better off isolated in their own communities, carefully shielded from opinions they do not agree with, instead praising TikTok’s ultimate filter bubble enabled by an almighty algorithm. He simply ignores all past instances where algorithms promoted hate speech, disinformation, propaganda, conspiracy theories, helped influence elections. In the most recent example, 2020 exam results in the UK were generated by an algorithm, leading to accusations of bias and unfairness. The article contains no less than 40 mentions of the word, as if Bytedance suddenly invented the panacea for every global problem. A world where everyone is happily isolated from reality, passively consuming amusing video shorts carefully selected to increase their addiction sounds rather dystopian to me. It also makes people less critical and more susceptible to hidden propaganda. Having this inescapable filter bubble controlled by the Chinese government is precisely the danger Ben Thompson is warning about in his article.

For all the naïve and idealistic dreams of the so-called “marketplace of ideas”, the first generation of large social networks has proven mostly unprepared and ill-equipped to deal with the resulting culture wars. Until they have some real substantial ideas and incentives to take on the costly task of mediating between strangers who disagree with each other, they’re better off sorting those people apart. The only types of people who enjoy being thrown into a gladiatorial online arena together with those they disagree with seem to be trolls, who benefit asymmetrically from the resultant violence.

Consider Twitter's content moderation problems. How much of that results from throwing liberals and conservatives together in a timeline together? Twitter employees speak often about wanting to improve public discourse, but they’d be much better off (and society, too) keeping the Slytherins and Gryffindors apart until they have some real substantive ideas to solve the problem of low trust conversation.

Despite this incessant fawning around its algorithm, I do not see much value in TikTok for the long term. Microsoft may want to get access to this miraculous algorithm, but short video apps have come and gone multiple times, with many alternatives operating now. The same is true of social media in general, younger generations adopting new online spaces and forms of expression to differentiate themselves from earlier generations. As Wei notes in the article, the popularity of TikTok is at least partially built on massive ad spending and on people downloading clips and sharing them to Instagram stories. It would be trivial for Facebook to stop accepting their ads and to block reuploads of TikTok clips, stunting their growth – but under the current antitrust scrutiny, not something they can do overtly.

Post a Comment