The uncanny accuracy of recommendation engines and what it says about us — and the companies watching.
It starts with a video about grief you didn't ask for. Then another. Then a thread about estrangement, a poem about loneliness, a comedy sketch that somehow nails exactly what you've been feeling for six months but couldn't name. You didn't search for any of this. The algorithm found you.
Recommendation engines have become eerily good at emotional profiling. Not because they understand you — they don't — but because they've processed enough human behavior to recognize patterns you might not even see in yourself. The content you pause on for two extra seconds. The videos you rewatch. The comments you type and delete.
Therapists spend months building the kind of picture these systems construct in days. And unlike a therapist, the algorithm never forgets, never gets tired, and never has competing interests in your wellbeing. It has one interest: keeping you on the platform.
That's the uncomfortable truth underneath the uncanny accuracy. The algorithm isn't trying to help you. It's trying to hold you. The emotional resonance is a byproduct of engagement optimization, not care. Every time it gets you right, it's because getting you right keeps you scrolling.
Which raises a question worth sitting with: if a system that doesn't care about you can understand you this well, what does that say about the systems — and people — that do?
Liam Park
Staff writer at The Daily Lens. Covering culture, ideas, and the stories that shape how we live.
