Oh, HERE's what I've been doing wrong:

Not killing enough experiments.  Especially with my Android apps, I'm not very comfortable with unpublishing them or otherwise letting them die.  I want them to be in my "portfolio", so that when people want to know if I'm good at programming or research or whatever, I can say "yes look at these things that I've made."  The irony is, if I just killed the ones that end up being hard, useless, and joyless, I could probably have made a lot more cooler things by now.

Well, or maybe not.  Maybe the point is I'm trying too hard.  Maybe I've set goals that are too high.  (get into grad school by wowing everyone with mad sick hackery and profound thoughts aplenty while I keep working a regular job; nay, a demanding intense job that also requires said hackery and thinkery?)  Maybe I shouldn't be setting goals at all!  Well now that sounds profound and profane and provocative, so it must be right.

No, but seriously, though.  I feel a little lighter every time I stop thinking about end results.  Whenever I stop worrying about whether I'll get into grad school and start reading more papers or exploring more whatever in my spare time.  I get bouts of "whatever happens, I'll probably do pretty okay", which really takes the weight off, and lets me live by curiosity that I can explore at will instead of goals that I need to grind out.  It's a really sane, sustainable world view.  (then it all wears off and I worry about success and failure again, but what can you do.)

I'm sure there's a good case to be made for goals.  And y'know, everything in moderation, and sometimes goals are good, yeah yeah.  But consider, maybe there's a good case to be made against goals too.  I never really thought about that.