I'm realizing that most of game development isn't actually development, but improvement. An example is Immortal Defense -- I basically made the entire game in the first six weeks, and after that it was just adding nice features, polishing, improving, finishing resources (graphics and music), balancing, playtesting, removing bugs -- that stuff took 80% of the work, actually *making* the game (the part that I most enjoyed) was only the first 20%. But the other 80% was necessary, if tedious. Not totally unenjoyable but not as enjoyable, and usually more frustrating.

Now that I'm working on another game (Saturated Dreamers), I realize again how much more fun that first 20% is than that last 80%. I'm envious of people like Miyamoto and Kojima, who only really spend time on that first 20% and have other people do the other 80%. But the sacrifice they make in exchange is not having a hand in every little detail of a game, so each has its advantages.