There's a moment — and you've had it — where someone you love hands you a book, or puts on a song, or orders for you at a restaurant, and they just nail it. Not because they ran the numbers. Because they know you.
They know you hate cilantro but love lemongrass. They know you cry at movies about fathers and daughters. They know you're in a weird mood this week and need something light, something that won't ask too much of you.
No algorithm has ever done that.
The Engagement Trap
Here's what recommendation engines actually optimize for: time on platform. Not satisfaction. Not delight. Not the quiet thrill of discovering something that changes how you see the world. They optimize for not leaving.
This is why Netflix recommends the show you'll half-watch while scrolling your phone. Why Spotify's Discover Weekly is 80% fine and 20% songs you already know by artists you already follow. Why Amazon thinks buying a toilet seat means you want to see forty more toilet seats.
These systems aren't broken. They're working exactly as designed. They just aren't designed for you. They're designed for engagement metrics, retention curves, and quarterly earnings calls.
The result is a strange kind of abundance that feels like scarcity. You have access to every book, every album, every film, every restaurant — and you spend twenty minutes scrolling before giving up and rewatching The Office again.
What Humans Do Differently
When your friend recommends something, they're doing something no algorithm can replicate: they're modeling your inner life.
They're not looking at what you clicked on. They're thinking about who you are. Your sense of humor. What you're going through right now. That conversation you had last week about feeling stuck. The thing you said once about loving stories where the quiet character turns out to be the hero.
This is what psychologists call mentalizing — the ability to hold a model of another person's mind inside your own. It's one of the most sophisticated things the human brain does. It's also, not coincidentally, what makes a recommendation feel like a gift instead of an ad.
A friend doesn't say "users who bought X also bought Y." A friend says "I know you're going to love this, and here's why." The why is everything. It's context. It's care. It's proof that someone was thinking about you when you weren't in the room.
The Trust Economy
Think about the people whose taste you actually trust. Not influencers with affiliate links. Not critics performing objectivity. The real ones — the friend who got you into that band in college, the coworker whose restaurant picks never miss, the person on Letterboxd whose four-star reviews are more reliable than anyone else's five.
You trust them because you've been calibrated against each other. You've disagreed enough to know where you diverge, and agreed enough to know where you align. That calibration takes time, shared experience, and vulnerability. You can't fake it with a cosine similarity score.
This is the economy that matters. Not attention. Not engagement. Trust.
When someone with great taste puts something on their shelf and says "this one's important," that signal cuts through every algorithm, every trending list, every "you might also like." Because it carries weight. It carries a person's reputation, their identity, their taste — all the things they've built over years of paying attention to what moves them.
Curation as Care
There's a word we keep coming back to at Stacks: curation. It gets thrown around a lot — curated playlists, curated feeds, curated collections. Usually it just means "we picked some stuff." But real curation is something deeper.
A museum curator doesn't just pick good art. They create relationships between pieces. They build a journey. They make an argument about what matters and why. They put things next to each other that have never been next to each other before, and suddenly both things mean more.
That's what the best recommenders in your life do. They don't just say "this is good." They say "this is good, and it connects to that thing you already love, and here's why both of them matter." They curate for you, and in doing so, they show you something about yourself you didn't know.
This is an act of care. It takes effort and attention and a genuine interest in someone else's experience. It's the opposite of an algorithm, which takes nothing and gives you more of the same.
The Shelf as Identity
Here's something interesting about Letterboxd, Goodreads, and every collection app people actually love: the profile page matters more than the feed.
People spend real time arranging their shelves. Choosing what goes in the top four. Deciding whether to rate something four or five stars. Writing a review that says something real. They do this because a shelf isn't just a list — it's a self-portrait. It's you saying: this is what I care about. This is what I've noticed. This is what I think is beautiful.
And when you visit someone else's shelf, you're not consuming content. You're meeting a person. You're learning what they love and how they see the world. The discovery that follows — "oh, you like that? I need to try it" — doesn't feel like a recommendation. It feels like a conversation.
Building for Humans
We built Stacks because we believe the best recommendations will always come from people, not machines. Not because algorithms are bad — they're fine for some things. But for the stuff that matters? The book that changes your life, the album you play at your wedding, the restaurant where you have the first date that actually goes somewhere?
For those, you need a person. Someone who knows you. Someone whose taste you trust. Someone who put something on their shelf and, without saying a word, told you: this one's for you.
The algorithm doesn't know you. But your people do.
Stacks is a social curation app where your taste speaks for itself. Coming soon.