Stop letting the algorithm choose what you watch

You open an app to watch one thing. Forty minutes later you are still there, watching something you would never have searched for, with no clear memory of deciding to. The video was fine. That is the point. It held you, which is the only thing it was selected to do.

This is not a personal failing, and it is not really about any single app. It is about what recommendation engines are built to optimize, which is almost never the same as what you actually want.

Engagement is the only thing being measured

A recommendation system does not know your taste. It knows your behavior, which is a much narrower and less flattering thing. It watches what you click, how long you stay, and what you do right before you close the app, and it tunes itself to produce more of whatever kept you there. Taste is what you would choose with a clear head. Engagement is what you do at 11pm when you are tired and your guard is down.

Over time, the gap between those two compounds. The feed is not learning who you are. It is learning which version of you stays longest, and then it shows that version more of itself. This is true across formats: video apps, music apps that auto-queue the next track, social feeds, even shopping apps that decide what you see before you have decided what you want.

The quiet costs

None of this announces itself. The costs are subtle, which is exactly why they accumulate. A few worth naming:

• Your world narrows. The feed keeps serving variations on what already worked, so the genuinely new and the slightly difficult rarely surface.
• You consume a lot but choose almost none of it. At the end you can describe what you watched, but not why you watched it.
• The feed learns your weak moments. It knows when you are bored, restless, or avoiding something, and it has an answer ready for each one.
• Time stops feeling like a decision. There is always a next thing, and saying no requires effort that watching does not.

If you want the longer version of why this happens, the design incentives behind it are worth understanding: why feeds are designed to never end.

Recommended to, versus choosing on purpose

There is a real difference between being recommended to and choosing something. When you are recommended to, the app holds the steering wheel and you approve its turns one at a time. When you choose, you decide first and the app fetches what you asked for. Both can fill an evening. Only one of them is yours.

The good news is that choosing on purpose does not require willpower or a digital detox. It mostly requires using the parts of these apps that were already there before the feed took over: search, subscriptions, playlists, and the quiet recommendations of actual people.

Taking back the steering wheel

You do not have to quit anything. You just have to do a little more of the deciding. A short list that works:

• Arrive with a query. Open the app already knowing what you came for, and search for it instead of scrolling toward it.
• Lean on subscriptions and playlists. A list you built is a decision you already made, calmly, in advance.
• Trust people over engines. A thing a friend sent you is recommended by someone who knows you, not by a system that knows your weak moments.
• Save instead of consuming now. When something interesting appears, save it for later. Most of the urgency is manufactured.
• Remove the feed where you can. If you mostly use video for one thing, an app that hides recommendations entirely takes the decision out of your hands in the good direction. talavo is a free, distraction-free video browser whose Zen Mode hides recommendations, feeds, and comments, so the only thing left to do is the thing you came to do.

If your reason for opening these apps is to learn something specific, it is worth being deliberate about that too: using video to learn without the rabbit hole.

The small dignity of deciding for yourself

None of this is about discipline or self-improvement. It is about a small and ordinary kind of dignity: the difference between an evening that happened to you and one you chose. The algorithm is very good at filling time. It has no opinion about whether the time was worth filling. That part was always yours, and it still is. You just have to take it back, one decision at a time, starting with the next thing you open.