What highly controversial 'heating' practice did employees at major short-form video apps use to manipulate culture?

The hidden hand behind a viral video

On a short-video app, a clip can look as if it caught fire purely because viewers loved it. But employees at major platforms have also used an internal practice called “heating”: manually giving selected videos extra exposure in recommendation feeds. Instead of waiting for the system to infer that a clip deserves a wider audience, a worker can help put it in front of more people.

That matters because recommendation feeds are usually built to learn from behavior. A platform shows a video to some viewers, measures signals such as whether they watch or skip, and uses those signals to decide who might want to see it next. Heating adds a human-directed push to that process. It can make a video seem to have gained momentum organically, even though an employee helped start that momentum.

Why platforms would do it

A recommendation system is good at predicting what may hold a viewer’s attention, but prediction is not the same as editorial judgment. Staff may want to spotlight a creator, introduce a new feature, or make sure a particular kind of video gets seen. Heating offers a direct way to influence what enters the feed without changing the underlying algorithm for everyone.

The tradeoff is control versus trust. A platform can use the practice to shape what users encounter, but viewers may assume that a widely circulated clip rose there solely through audience response. If the boost is not clearly distinguished from ordinary recommendations, the feed’s apparent popularity signals become harder to interpret.

When a boost becomes culture-shaping

Extra exposure can set off a feedback loop: more people see a clip, some react or share it, and those reactions can draw still more attention. A small editorial choice can therefore affect which sounds, creators, jokes, or topics become familiar across the app. Heating does not guarantee a hit, but it can give selected content a head start that ordinary creators may not receive.

The consequential question is not only which video gets a boost. It is whether users can tell when the platform itself helped decide what became popular.