I’ve found AI to be incredibly useful for getting things done that otherwise would not get done.
For instance, an important-but-not-vital feature was blocked by the need to do a huge refactor - to replace a complex runtime system with thousands of lines of source generators, producing tens of thousands of lines of parsing code. That daunting task sat on my backlog for years.
Then one day, I chuckled to myself, “I wonder what those dumb robots would do with this!” A few days later, it was finished, and the feature could move forward.
When I had a fun idea that required some media with high production value in order to come out as I envisioned it, I could, theoretically, have scoured the internet or library archives for the exact type of 1980s promotional photo I had in mind, and hoped to find one with high enough resolution, no watermarks or licensing restrictions, etc. Or I could, theoretically, have hired a handful of actors and a professional photographer to create it from scratch.
But realistically, that wasn’t going to happen - this was a project I was doing for free, and there was already a lot of work involved in the rest of it.
So instead, I typed in a prompt for exactly the image I wanted, did some photo editing to artificially age the result, and that part of the project was done.
When a throwaway remark someone made on Discord gave me a speculative idea that required writing a bunch of timing-sensitive assembly code for a processor I knew nothing about, I could, theoretically, have devoted the next several weeks of my life to learning the ins and outs of that chip and its development ecosystem, then iterated for who knows how long just to learn whether the thing I had in mind was even possible.
But realistically, I would just have said “huh, I wonder if you could do X” and shrugged it off. So instead, I spent a fraction of what I had paid for lunch that day on AI credits, and in a couple hours, I had a working demo.
When I caught wind that one of my projects was going to be featured in an announcement from a big company, I wanted to have a slick-looking website ready for the influx of interest. I could have frantically tried to learn everything I could about the last 20 years of web development, and slapped together something that looked okay on some browsers some of the time. Instead… well, you get the picture. I now have not just a slick-looking website, but a slick-looking web application that lets people try it out without even downloading it.
I like programming. I’ve been doing it for fun since I was 4 years old, and as a career for decades. But it turns out I also like sketching out plans for programming, watching someone something implement those plans and kibitzing while they do it, and having a finished product I can tweak by hand as needed.