I do this thing every time I load a roll into the Leica that I’ve never fully figured out. Before I close the back, I run a finger along the film and feel it pull tight. Half a second, maybe. But by then I already know the next thirty-six photos are going to be nothing like whatever I could make with the phone sitting in my other pocket.
I wanted to write about this precisely because it has nothing to do with my actual life. I spend my days building things with AI, thinking about how machines can do more for us, faster, better. And then I went and bought a camera from 1984 that does absolutely nothing for me. That contradiction felt worth sitting with.
A camera that doesn’t help
The M6 doesn’t help. No autofocus, no auto mode, no screen to check what you got. It’s you and three decisions: aperture, shutter speed, focus. You read the light off a tiny meter inside the viewfinder, you turn the focus ring until the two images in the rangefinder snap into one, and only then, when everything is where it should be, you take the shot.
It’s awkward. That’s the whole point, though. Every frame forces you to be there. You can’t hand anything off to the camera because the camera can’t do anything on its own. You look longer, you wait more, you think before you press. After a while you start seeing light differently. You notice the way it hits a wall at six in the evening. You start doing the math without realizing it. The camera doesn’t solve the photo for you. It trains you to see it.
And then there’s the sound.
That click
The shutter on an M6 makes a noise that’s hard to describe to anyone who hasn’t heard it. Dry, mechanical, short. A clean little metallic snap that lasts a fraction of a second. Nothing simulated, no effect layered on top. It’s an actual cloth curtain opening and closing, exposing film to light for exactly as long as you decided it should.
That click is final. The moment you hear it the photo is already gone. You can’t see it, can’t delete it, can’t shoot it again the same way. It’s now sitting on a strip of plastic you’ll develop in two weeks, by which point you’ll have half-forgotten you took it. There’s something honest in that. The photo either exists or it doesn’t, and you put yourself on the line in that one instant.
Thirty-six frames. Not thousands. Thirty-six. Each one costs something, each one counts, and that scarcity changes everything. You don’t spray and pray. You wait for the scene to be worth the frame.
What the phone does (and does brilliantly)
In fairness, because this is where it gets interesting: pull out the phone and point it at the same scene. Three or four cameras firing at once, grabbing several exposures in parallel, fusing them with models that decide the dynamic range for you, the noise, the contrast, the white balance, even where to put the focus after you’ve already pressed the button. The result is technically flawless. Sharp, even, “pro” without you knowing the first thing about photography.
I mean that without any sarcasm. It’s a feat. It’s one of the most quietly impressive pieces of applied AI we have, tucked into a thing we carry around all day. As someone who builds this stuff, I admire it for real. Computational photography is genuinely brilliant.
But there’s one thing it can’t do, and it has nothing to do with quality.
The part the machine can’t make
The phone hands you a perfect image. The Leica hands you an experience. That’s the whole difference, and I think it gets lost when the argument collapses into “film just looks better” or “grain has soul.” That’s not it, or not only it. What makes a film photograph what it is doesn’t live entirely in the result. It lives in what happens to you while you make it. The phone optimizes the output and erases the process. The Leica does the opposite. The process is the thing.
Here’s what I keep coming back to, and it’s a strange thing to believe while spending all day on the other side of it: the better the machine gets, the more valuable the stuff we choose to do slowly and by hand becomes. Not because the machine is bad. Because it’s good. It’s exactly because the machine can do it perfectly that the human act gets freed up to be about something else. Being there. The decision. The body holding the camera, the finger metering the light, the ear waiting for that click. There’s a kind of satisfaction in doing things by hand that isn’t nostalgia and isn’t a protest against progress. It’s older and dumber than that. Friction, when you choose it, gives the gesture its meaning back.
AI can already make a better image than I can. That’s true now and it’ll be truer every year. What it can’t make is the feeling of loading the roll, reading the light off an afternoon in Barcelona, waiting for the moment, hearing that shutter, and knowing something got caught for good in an instant that isn’t coming back.
Islands worth keeping
I don’t think this is really about photography. There’s plenty we’re going to be able to hand off to AI, and we should. But there will be these islands of deeply human activity that make no sense to optimize, because the whole value of them is in doing them ourselves. Slowly. Badly, sometimes. With your whole body in it. Cooking. Playing an instrument. Writing by hand. Shooting thirty-six photos in a day and waiting two weeks to see them.
We don’t do those things because they’re efficient. We do them because they make us feel alive.
Tomorrow I’ll load another roll. I’ll run a finger along the film and feel it pull tight, and somewhere in the afternoon I’ll press the shutter and hear that dry, perfect click.
It’s the least technological thing I do. And in a world full of machines that think, it’s one of the things I care about most.