From Film to Feed: A Social Media Experiment with AI
If you’ve been listening to the Patreon Specials, you probably know about the experiment I’ve been running with my personal Instagram account (@_paul_u). It just wrapped up, and here’s the story:
For an entire year, I didn’t personally post any of the 550+ photos on my Instagram account… or at least, not directly. Let me explain: A friend helped me set up a small piece of code that we hosted on a server, along with a collection of photos I took over the past few years. These included good shots, bad ones, random ones—pretty much everything from terrible snaps to print-worthy images.
Each picture would go through a combination of Google Lens, ChatGPT, and other tools that I honestly don’t fully understand, to generate a caption made up of one to three emojis, depending on what was in the picture or the general “color” of the subject. The code would then automatically add the hashtag #filmisnotdead and post the photo whenever it deemed appropriate, with the goal of uploading the entire gallery by the end of the year. We did also add color-theory to post the pictures in a nice gradient rather than in a random order.
As for the emojis, the system mostly worked, often creating interesting patterns. Every time it uploaded a picture with motion blur, the “fast” 💨 emoji would accompany it. Sunset photos were mostly tagged with the “Orange” 🍊 emoji due to the color, but over time, it started using the more appropriate “Sunset” 🌅 emoji. I occasionally added some hashtags, locations, or tags here and there, but I tried to stay as hands-off as possible.
So, why did I do this in the first place? A lot of social media experts will tell you that “CONSISTENCY” is the secret to growing your following. If you’re a professional or a brand, increasing your social media following can be a great way to attract more clients, etc. So why doing it with my personal account? The answer is simple: I can’t plan a year’s worth of content in advance on Slow Spin Society, and the end goal was more about seeing what would happen rather than gaining followers—which I don’t really care about.
Algorithms and other 21st-century atrocities aside, what did I personally learn from this whole experiment? Well, as someone in photography and media creation, I often feel pressure to put new content online. “What if this picture gets me a new client or an opening in a gallery?” The problem is, I’m also incredibly lazy and don’t want to spend time on social media, so I came up with this scheme to relieve that pressure and reduce my screen time. But here’s the irony: shooting on film, only to ultimately post it on a platform where people immediately forget what they saw on their tiny screens. And using AI to do so, is just as crazy as thinking that using social media to find serious opportunities is a viable strategy.
So what now? Well, I’m going to do what I should have done with social media all along: not take it too seriously. I’ll upload organically when I feel like it, without feeling pressured if there isn’t anything new for two or three weeks. Sure, I could say the ultimate goal would be to ditch Instagram altogether, but I enjoy seeing what my friends are up to sometimes. There’s nothing inherently wrong with the concept of social media—the problem lies more with the companies that run them, serving you ads by the ladle, selling your data, and using your content to feed more AI. It’s fine to have an online diary of what you’ve been up to, as long as it doesn’t overshadow what you’re actually doing—or not doing—in real life.
What experiment should I run in 2025?
🎞: Many
📷: Many
📍: Many
