My daughter is about to turn 15 and a half, the age when you can get a driver’s license in the US. But since she’s still under 18, she first has to complete about 30 hours of online/offline driver’s education and get a certificate. Turns out there are several platforms that offer the course online, so we subscribed to one and she’s been taking the classes. She must have reached the chapter on parking, because a few days ago she asked me this:
“Why do you have to back into a parallel parking spot? Can’t you do it going forward?”
“If there’s enough room, nothing stops you from doing it going forward. But usually you don’t have that much space.”
That conversation suddenly reminded me of the mini games from the days when Flash ruled the web, so I sent a quick set of requirements to my agent team and had them build one. As it happens, I had just set up Envy, the designer, the day before. So I told Breda, the developer, to build it and to get the design assets from Envy. I sent the request and went off to do other things, so I didn’t see the in-between versions. I think there was one version that only tested the mechanics before the design assets arrived, and the version with the assets came right after. Time taken: about 25 minutes. It worked well, so I asked for more stages, and then for an English version. Each took about 10 minutes, I think.
These days it’s hardly exciting anymore when AI whips up something like this. What makes this iteration meaningful to me, though, is that it’s the first cycle where a single task in the issue tracker took us all the way through coding, design asset creation, GitHub repo creation, and online deployment (Google Cloud Run). It means all the skills and permissions this cycle needs are in place. Of course, I don’t have much of an eye for design, so the assets are pretty crude for now, but that can be improved over time. Maybe I’ll build a few more of these with the team, just for fun.
