I nodded along to a piece of advice this summer that took me about a month to actually hear.

It came during ESR, a research program I was part of, from Dr. James Duban, who has spent decades at the University of North Texas guiding students through research and national scholarship competitions. We were talking about mentorship, and he told us not to work with a mentor for the letter of recommendation, and not for the publication either. Work with them to learn from them, he said. And then find a way to give back.

At the time this registered as standard-issue mentorship advice, the kind of thing you nod at because it sounds nice and costs nothing. (I am very good at nodding at advice that costs nothing.) It took me until roughly the end of the program to notice that it's actually a radical claim, because it inverts the entire logic by which my generation has been taught to build relationships.

The standard model, the one implicit in every pre-professional club and LinkedIn success post, treats a connection as an asset: You meet someone because they might write you a letter, refer you to a program, or introduce you to someone who can. The relationship is a door, and the question you're trained to ask (quietly, politely, but always) is what's on the other side of it. Duban's advice cuts against this at the root. He wasn't saying the letter doesn't matter. He was saying the letter is downstream of something else, and that if you aim at the letter, you'll miss the something else entirely.

But the more interesting question here isn't "whether networking is good or bad." It's what happens to a relationship once it becomes a means to an end.

Aristotle worked this out about 2,300 years before LinkedIn did. The Nicomachean Ethics distinguishes three kinds of friendship: friendships of utility, friendships of pleasure, and friendships of virtue. The first two are instrumental—you value the person for what they provide, whether that's a favor or a good time. The third is the one where you value the person as the person they are. And his observation about the instrumental kinds is the load-bearing part: They dissolve the moment the utility does. When someone can no longer provide the thing, the friendship evaporates, because the friendship was never with them. It was with the thing.

This is, I think, a precise description of what feels off about the culture forming around me. Everyone calls each other for favors. Every coffee chat has an agenda tucked somewhere behind it. Every hangout is secretly a way into some room, some club, some list. And I notice that nobody says any of this out loud, which is itself telling: If instrumental connection were actually fine, we wouldn't need to disguise it as friendship. The disguise is a confession.

What strikes me most about the favor-trading equilibrium isn't that it's cynical. It's that it's ungrateful, in a specific structural way. Gratitude requires receiving something as a gift: something you weren't owed, given by someone who didn't have to give it. But in a system where every interaction is an implicit exchange, nothing is a gift. The mentor's time isn't generosity; it's an investment to be repaid in acknowledgment-section credit. The introduction isn't kindness; it's a favor to be banked and called in later. You cannot be grateful inside a ledger. You can only be even or in debt.

And this is where the second half of Duban's advice (give back) stops sounding like a platitude and starts doing real work. Giving back isn't repayment. Repayment closes a transaction; giving back continues a relationship. When he told us to give back to our mentors, he wasn't describing the other half of a trade. He was describing what a relationship looks like when neither side is keeping score—when the thing being exchanged isn't favors but attention, care, and the particular pleasure of learning from someone who knows more than you do.

Here's the strange part, and the reason I don't think this is just sentimentality: The instrumental approach fails even on its own terms.

Two observations. First, the best letters of recommendation are written by people who actually know you, and people only actually know you if you showed up for reasons other than the letter. A mentor can smell a transcript-builder from across the lab, and the real investment (the unprompted phone call to a colleague, the honest edit, the opportunity you didn't know existed) gets reserved for students who seem to be there for the work. Second, the network that actually catches you when you fall is made of people who like you, not people who owe you. Debts get discharged. Affection compounds.

So even the pure careerist, if they were being rigorous about it, should behave exactly as if connection were the end rather than the means. The instrumentalist's optimal strategy is to stop being an instrumentalist. (Whether you can fake your way into that equilibrium is a separate question. My guess is mostly no; people are better at detecting instrumentality than we are at performing its absence.)

To be clear, none of this adds up to "don't network."¹ Meet people. Ask for things; let people ask you for things. The doors matter, especially for those of us who didn't inherit keys. The claim is narrower, and it's about the order of operations: The letter, the referral, the room are byproducts of relationships, and byproducts have the inconvenient property that you can't get them by aiming at them directly. Aim at the person. Learn from them because learning from them is good. Give back because you're grateful, not because you're settling accounts. The doors mostly open anyway.

The hard part is that I can't fully prove any of this. Somewhere out there is a ruthlessly transactional networker doing just fine, and plenty of generous people go unrewarded. But when I think back on ESR, the part I keep returning to isn't the research. It's that I got to learn from someone who has spent his career doing exactly what he told us to do: teaching students, and giving back. I'm grateful I was in the room for it, even if it took me a month to hear.