The "Good Enough" Trap

The market is saturated with AI-generated work that looks almost right. Almost polished. Almost on brand. Good enough to ship, not good enough to remember.

There are two reasons this is happening, and they're worth separating.

The first is budget and time. AI generation is iterative - getting from "close" to "right" costs tokens, and tokens cost money. I recently needed concept art for a character. I described exactly what I had in mind, got something close but not quite there, and made a deliberate call to stop iterating. For a personal project, that tradeoff was acceptable. For client work, it wouldn't have been. When budgets are tight and timelines are tighter, "good enough" stops being a compromise and starts being a policy.

The second cause is harder to fix. Generating output with AI is now genuinely accessible to almost anyone. Knowing what good looks like is not. Taste, judgment, and recognizing when something is off come from experience, not from access to a tool. When that experience isn't in the room, no amount of iteration closes the gap.

Both problems produce the same result but they need different solutions. The first is a strategy problem - knowing which tasks AI handles well, which it doesn't, and where the iteration cost stops making sense. The second one has been around longer than AI. Back in the early 2000s, "everyone" could suddenly build a website (spoiler alert - none of those websites survived the test of time).