Try a Moon Shot Idea Every Day

I subconsciously started doing something a while back, and now it’s become an intentional habit. In a side screen, I’ll give Claude an impossible task. Refactor something huge, or try to one-shot an application with a big prompt. The success rate is somewhere in the 5-10% range. That might seem low, but it’s essentially a marginal-cost zero half court shot that pays big when it works. This is largely a side effect of delayed market effects on an industry gobbling up capital and trying to convert long term loyalty through unsustainable subscription prices. As such, I don’t think it will be around forever. While it is, might as well get some big wins with minimal marginal effort.