Stop calling product recommendations "personalized"

"Personalized" recommendations aren't personalized to you, they're personalized to people who behave like you. Every product feed calls itself personalized now, but look at how most of these systems actually work. Collaborative filtering, still the backbone of most recommendation systems, works by finding other users with similar behavior patterns and showing you what worked for them. It's modeling a cluster you've been sorted into. It doesn't actually know who you are or what you want. It might work occasionally if it's recommending books or movies, where taste does tend to cluster by demographic and behavior. But two people can behave identically online (same clicks/searches/purchase history) and still need completely different things because of their personal experiences, culture, net worth, appearance, etc. So when a "personalized" system gets something wrong in one of those categories, is it actually broken, or is it working exactly as designed and the category itself needed something other than behavioral clustering? It really seems like devs are getting lazy with AI, especially since it looks like the same recommendation architecture keeps getting applied across categories with fundamentally different requirements.