I have enjoyed making images for my blog posts with ChatGPT. Yet, I was thinking about it, and I fully expect that producing those images takes significant amounts of water and electricity. So I am going to draw stuff myself. It doesn’t seem right to participate in the catalysis of water shortages just so I can chuckle about a nonsensical image I wanted ChatGPT to make.
Moreover, I get numerous benefits out of producing the artwork myself (I’m using a verrrrry loose definition of the word “art” here). For instance, by drawing, myself, I put the onus of color selection on me, rather than allowing a computer to decide for me. Personally, I prefer to be — as GWB put it — “the decider.”
I sincerely like automation. I love being able to simply flip a switch then get light in the dark. However, I prefer to automate tasks I dislike. It makes no sense to me to automate the tasks I enjoy, and I sincerely like drawing and exercising creativity. I like starting out with a few fragments of ideas and seeing them coalesce and form a whole which I had previously not envisioned. I like individually participating in the production, and I like working with flaws and imperfections. This notion that LLMs are good because they remove friction is kinda missing the point. Sure, if we can automate away undesirable aspects of processes, I’m for it. However, a lot of friction is functional and closely related to joy. I don’t want to eliminate _all_ friction, as much as marketers want to convince people all friction is bad. Like, without friction, fires would be _a lot_ harder to light. Dumbass.