Thirteen minutes of light
I added up every second my camera’s shutter has ever been open. Then I did it for all of humanity.
A few weeks ago I started writing an email to an old professor of mine from Tisch. I was going to say something about how the time since graduating feels like forever and also like nothing at all… and then I got sidetracked (as usual) by a different question. How much time have I actually captured?
Not how many photos. How much time. Every photo is a tiny slice of the light visible in a certain amount of time, so if you add up the slices from every frame I’ve ever shot, you could technically get the total amount of time my shutter has been open. My guess was hours. Maybe days across someone’s entire career. It was actually about 13 and a half minutes. Part of me thought that was lame, but it was also kind of cool to think that in spite of the years and years I have spent on the craft, the amount of time that “mattered” was genuinely tiny. I feel like that representation is a good way to think about the idea of “time flying”.
How I counted
Every digital photo records its shutter speed in its metadata, so I had an agent read it from every raw file on my drives, throw out the duplicates, and add it up. It found roughly 38,000 photos from 2022 on. If I keep shooting at this pace until I retire, I’ll land somewhere between one and two hours… my entire career as a photographer might fit inside the time it takes to set up and break down lights on set.
Before I ran anything, I guessed I probably shoot around 1/125 of a second. I was kinda right, that was my most commonly used shutter speed. But my average is 1/46. Slower than how I’d shoot in most scenarios, and the reason is kinda interesting…
About 2% of my frames take up 70% of the time. Twenty-nine long exposures add up to nearly as much open shutter as the 7,467 frames I shot between 1/125 and 1/250.
The slowest 2% of frames hold 70% of the time
- 1/125 or faster
- 1/125 to 1/15
- 1/15 to 1 second
- 1 to 30 seconds
The stuff I think of as my work (sharp, fast, well lit) barely registers. The time is filled and skewed by long exposures.
Ten thousand years of light
I didn’t know this off the top of my head, but probably learned about it in art history or something and forgot. I made a little widget that puts fun facts and quotes on my MacBook header and thought this one was cool: The first photograph, taken by Nicéphore Niépce in the 1820s, needed an exposure of at least eight hours.
So then I scaled it all the way up. Common estimates put the total number of photos ever taken at 13 to 15 trillion, with about 2 trillion more every year now, almost all on phones. Using some rough assumptions about the history of imaging processes (ex: how fast people shot in each era, from glass plates to roll film to digital), all of humanity’s open shutter comes to something in the realm of 10,000 years. Obviously this is a major estimate, not a measurement. But if we assume I’m right about everything: Per person, humanity has captured a smaller share of its existence than I have of my own.
Years of open shutter captured worldwide, per year
in 2025 alone About half of all the open shutter
ever captured came after 2020
World population
Cost of one photo, in 2025 dollars
Photos per person, per year log scale
Seconds of open shutter per person, per year log scale
Who has a camera in their pocket
54% of people own a smartphone (2023). About 76% have some kind of mobile subscription.
The archive is small and crooked
While I love photography, I think a lot of the time it’s fake. I mean it’s not a bad thing, it’s the artist’s nature to be sensationalist right? Who wants to be mediocre? If a photographer takes a photo of empty seats at a stadium, it’s negative; if they take a photo of the full seats at a stadium and avoid the empty ones, it’s glazing. If you somehow take a picture of everything, you’re a security camera. By that logic, there’s no such thing as an unbiased photo.
With video, I guess you could say that spoken words are harder to fake. Sure, there’s fiction and there’s nonfiction. But a combination of audio, multiple frames per second, etc gives more context. You are capturing a larger moment in time than photography. It is more difficult to “fake”.
In the universe we’re small creatures, and our documented existence is even smaller. And a lot of it is likely biased, erased, or completely manipulated (think the Persian Empire, which we mostly know through the Greeks who fought it and the Alexander who burned its capital). Bad documentation of history is a side effect of human instinct first, and of living in the third dimension second… a photo can only ever show one flat slice of one moment.
We’ve never recorded more. That’s not the same as recording more that matters. When everyone captures everything, footage is cheap and attention is the scarce part. A single photo from a time before history was recorded would be more valuable than everything we do have combined.
Slices of a fourth dimension
Gonna get a little abstract here, bear with me… When physicists imagine a fourth dimension, it kinda looks like a spread of everything in the past, present, future. We can’t actually see it, understand it, or come to terms with it; we haven’t even figured out how to be in two places at once, let alone everywhere at every point in time. We can assign that understanding to make ourselves feel better though. It’s kinda cool to think about how the more we photograph (freeze) snippets of time/light, the more complete our understanding of another dimension might be. You could say that it’s an actual representation of a fractional dimension… If you were to project that the amount of photos taken on earth keeps growing exponentially, as it has so far, we can assume that eventually we could document as far as the 3.5 with the documentation of light (the future makes this hard, but I’d predict that generative imaging might have an interesting role here. It’s predictive of most likely outcomes given proper training, so it could fill in the gaps between real frames. But a filled-in gap is a guess, not a record. It shows what probably happened, not what did. The more we lean on it, the more the 3.5 becomes a dimension we imagined instead of one we documented.)
2D photos of a 3D object
One cube, one moment, three angles. Head on it's a flat square. Turn it and the shading tells you it has depth.
3D moments of a 4D object
Same cube, three photos at three times. It moves, it turns, the light changes. Each photo is one slice.
Stack enough slices and the gaps fill in
Drag the slider to take more photos, then move across the trail to pull out a single frame. A few read as separate cubes. Enough of them read as one thing moving through time.
Turns out people have been trying to photograph the in-between for a long time. In 1878 Eadweard Muybridge lined up twelve cameras along a racetrack, each triggered by a tripwire, to figure out whether all four of a galloping horse’s hooves leave the ground at once (they do). A few years later Étienne-Jules Marey built a camera shaped like a rifle that shot twelve frames a second of birds in flight. Neither of them cared about a single moment. They cared about what happens between moments, which is pretty much the 3.5 I’m talking about.
And it’s already happening at a much bigger scale. In 2015, researchers at Google and the University of Washington pulled millions of random tourist photos off the internet and stitched them into time-lapses of places changing over years… glaciers retreating, buildings going up. Nobody took those photos to go together. A year earlier, a project called Scene Chronology did something similar with city scenes like Times Square, working out roughly when things in the frame appeared and disappeared. That’s what I mean when I say the more we photograph, the more of that extra dimension we fill in. Nobody is shooting for it. It just shows up once there are enough slices.
Between the frames
We pretty much know nothing and even if we had infinite time to research, our physical bodies prevent us from ever gathering the data we need to solve almost all of it.
Photography really hammered that perspective into me. When you’re shooting, you’re in a constant state of observation, waiting for something to happen (and honestly, it’s addicting). But with time and experience I’ve come to care more about what happens around the moment, not the moment itself. My friend Dylan and I talk about this a lot, enough that we want to make a short film about it.
Think about going to a party. It’s the reason you go, but the ride there and back is where you have the most meaningful conversation with your friends. In my opinion, the best parts of the night happen in motion. Same with anything risky… the butterflies hit before the bet lands or the date starts, not after. But you still need to go to the party for the “motion” to take place.
That’s kinda how I think about my thirteen and a half minutes now. Every frame is the party. It’s the reason I showed up with a camera. But the 3.5 lives on the ride there and back, in everything that happened between the slices I actually kept. I’ll never capture most of it, and neither will anyone else. Knowing that is what makes the few slices worth taking carefully.
Maybe I should start telling people that we use agentic fractional dimension visualization to produce media. That would be a sick buzzword to use at a networking event.