What Model Are You Using?
A question doing more work than it admits
I keep picturing two robots leaning against a sticky bar. The music is loud. The lighting is bad. One turns to the other and says, “Hey. What model are you using?” It’s meant to sound casual. It’s not. The question carries the same energy as “what’s your sign?” or “what’s your major?” A proxy dressed up as curiosity. A way of deciding whether this conversation is worth continuing before anyone gets too invested. We pretend it’s about capability, but mostly it’s about affiliation. About taste. About which arguments you’re tired of having and which ones you’re still willing to entertain.
What makes the question interesting isn’t that people care about models. Of course they do. What’s interesting is how quickly the model becomes a stand‑in for judgment, rigor, and seriousness. We’ve done this before. With languages. With frameworks. With schools. We know, intellectually, that the tool is rarely the point. And yet we keep asking, because it’s easier than talking about how someone thinks, how they iterate, or how they notice when an answer is wrong.
In practice, the difference shows up somewhere else entirely. It shows up in how people frame a problem before they ever ask a model anything. In how they notice when an answer is subtly wrong instead of confidently incomplete. In whether they iterate, constrain, and refine, or just accept the first fluent response and move on. Models matter, but they fade quickly into the background. What stays visible is judgment. Taste. The ability to use a tool without mistaking it for the work itself.
I can’t even imagine what replaces the question. Not because nothing will, but because whatever comes next will be even more implicit. Less legible. Harder to name out loud. We’ll stop asking about models the way we stopped asking about programming languages, or cloud providers, or operating systems. The differentiation won’t disappear. It will just move. Into how quickly someone can orient in a new problem. Into how they recover from a wrong turn. Into whether they can explain why an answer feels off before they can prove it. The question won’t go away. It will just stop sounding like a question.
I answer the question, because of course I do. And I know that depending on who’s asking, I’ve just signaled something. Taste. Alignment. Maybe even intent. I’m aware of the tiny recalibration that happens on the other side of the exchange, because I feel it too. In my head, the robots are still at the bar. One has answered. The other has nodded. Nothing explicit has been said, but the tone of the conversation has shifted. They’re not talking about models anymore. They’re deciding whether to stay, order another drink, or politely drift toward the door. The question did its job. It always does.
Alison + Wiggins