Gongzilla

The collapsing cost of dumb ideas

You know how in inspirational corporate trainings they always tell you that “there are no stupid ideas?” That’s only partly true. The problem with truly stupid ideas is that if you pursue them, you will be spending time and effort that could better be spent on a less stupid idea. The stupider your idea, the more you’re paying in opportunity cost. The cost of pursuing an idea, however, is elastic. With cheap artificial intelligence, it is now possible to pursue fantastically stupid ideas without regret.

I woke up way too early on a Saturday morning and to entertain myself decided to try the new o1 + canvas functionality in ChatGPT to create a little game. o1 was good at generating something in one shot, which is exactly what o1 is so good at, but after a few iterations I realised that I’m not loving the experience, and that meanwhile I am spending my weekly allowance of o1 requests, which I normally use for serious work, on nonsense. To continue, I copied the resulting code and pasted it into v0. I have no idea what models or techniques are used in v0, but I love this tool. It just works. It’s cool that ChatGPT now has a way for everyone to play around with interactive apps, but if you really want to have fun collaborating with AI on apps, v0 is where it’s at.

The resulting game is Gongzilla. It’s not particularly good, but it looks cool, and you can actually play it. I didn’t write a single line of code, and the entire process took less than an hour and about 20 iterations.

Some other things I created recently with v0 and other AI friends:

• Pop the Monkey! - classic retro game, especially fun to play on the phone

• Decisionator - rational decision-making tool

• Feed Explorer - tool for reviewing RSS feeds (useful if you like subscribing to obscure blogs and podcasts)