Seems to me that this is now exponentially true with AI coding assistants. If you don't understand what you're adding, and you're being clever - you can quickly

Seems to me that this is now exponentially true with AI coding assistants. If you don't understand what you're adding, and you're being clever - you can quickly end up in a situation where you can't reason effectively about your system. I'm seeing this on multiple fronts, and it's quickly becoming an unsustainable situation in some areas. I expect I'm not alone in this regard.