I’m spending a lot of time on the project, and enjoying a lot of that time. There have been some extended periods of frustration, though. It’s usually when I’m trying to accomplish something that I thought would be simple but it turns out otherwise. I had such an experience yesterday: I was trying to make an expanding card expand over the next row of cards without pushing them down. After many failed attempts and much research, I gave up and used a modal instead.
During these times of frustration and “stuckness,” I am easily distracted by other interesting amusements. Sitting at a computer, I have the ability to scratch at any curiosity, and that can be very tempting. Some distractions are useful–I am partway through two Udemy React courses and a text-based React course, and while each one strays far from the content I need right now, they are all tempting and fascinating in their own way.
Other things are probably less edifying. I have amused myself with AI image modification and generation. I’m fascinated by it, but it is a very tempting way to waste my time.
The first thing I did (of course) was to use it to make pictures of my family and myself. My results were very mixed–this is my son:
This picture looks a just like him–it was the first AI portrait I attempted, and it gave me great hope for the AI engine’s abilities. No subsequent portraits were nearly this successful.
For example, here is one of me. This is the closest of the ones I did to what I actually look like, but I don’t think it really looks like me.
But it looks more like me that the next one:
I think friends seeing this image would easily recognize my wife:
But to get there I produced a lot of pictures that I would not want to share with her.
So the portrait-ability of AI is somewhat suspect, though I notice it does celebrities very well. This is presumably because it has a rich source of reference material instead of working from a single photo. Where it is more interesting is in producing images that you thought of. This is especially useful when you want to show something you can’t find stock imagery of. I mentioned in a previous post that my oldest son, who is still grieving a dog who died young, imagined his dog in a field chasing butterflies. A purebred dog would probably be simple to make an image of, but his dog was a pit bull mixed with who knows what. I’ve come close a few times but I did better when I started with a dog and added butterflies later. The ones showing a dog chasing butterflies don’t look like him at all.
Here’s another one that I would surely have trouble finding in stock imagery. My oldest (the same one grieving the dog) has a tattoo quoting Camus: “One must imagine Sisyphus happy.” The first images I did were some of the best:
Unfortunately, they did not upscale well. They became less detailed, for some reason I don’t yet understand.
This project has consumed me for the moment–I am sure I would be better off if I distracted myself with exercise, or doing my taxes, or chopping wood, but when you’re already sitting at a computer the easy way out is right there. Fortunately my photography habit does lend itself to walking, at the least. In the past couple years I have taken simple joy in photographing the same places repeatedly over time. They are remarkable to me because they almost always look very ordinary when I’m looking at them, but when I go back through the photos I see great variety in the visions of the exact same spot. This gallery is an example of such a spot:
If you have read this far, thank you very much! If you are my teacher, thank you for making the non-project workload of this class so light. That enables me to spend my time on the project, which I think is the point.
Leave a Reply