Picking Impossible Things To Do

Sometimes I wish I've picked simpler games to play.

Trying to get good at software gardening, produce novel insights, resolve psychological hang ups, the whole business of being intrinsically interesting is all great.

Except if you want to achieve the impossible.

Every feat of art was once impossible. Swallowing 50 hotdogs in one minute, a backflip on a BMX, that funny way of high-jumping, climbing rocks without harness, folding a realistic origami horse. I heard Bruce Lee wanted to practice a kick so fast it can tear a paper in half without touching it. He didn't live long enough to get around to it.

What was once impossible only need to be done once for it to become the next bare minimum benchmark.

These are all domains where success is clearly defined and achievement is indisputable. I like that they have clean winning conditions that's not subjected whims, you did it or you didn't.

The price is that winners and losers are also well defined. Sometimes there's only one place for winners, and most people spend their life not being that.

I think of myself as an athlete working on software. But the nature of the work do not often give me the satisfaction of clear wins.

I emphasize "often" because it's complicated. The act of creating a machine that works is a clear win.

The more important aspect of doing it beautifully and maintaining it well is undefined. Any attempt to measure it gets Goodhart'd away.

Which brings me back to the question: what does it mean to achieve the impossible in such domains?

It possibly isn't even a valid question but still an important one. Breakthroughs in any field were once impossible, therefore all progress relies on figuring out what is impossible and have them invalidated.

I don't yet have a good framework on how to do that but I now suspect this is where close-minded low-curiousity people become useful. Listen to them say "X is impossible" and you've got your next quest for impossiblity laid out.