Dynamic Languages Flame War Fuel

Steve Yegge from Google posted a blog entry transcribing his recent talk given at Stanford. It's not necessarily formal, but really seems to capture what I've been believing more and more over the years. Computers and humans are different. They are so surprisingly different as to be mutually incomprehensible. Trying to model computers mentally is no longer possible. Or even necessary.

Because computers are in a very alien sense self-aware. And they understand our attempts to model their behavior better than we do. They are also unimaginably better at it because they understand their own "psychology". And we understand ours. But the two are mutually incompatible.

And computers are absolutely stunningly good at solving problems that our brains aren't even equipped for. Like calculating complex paths that aren't known until they've been traveled. Or determining the probability that a certain event will occur based on other observed events. Or doing absolutely nothing, then immediately doing something, and then doing nothing again. Without lapses in attention.

And since optimization of code is really just a mathematics problem, computers are great at it. So why try and outsmart them at it? They are provably better at it in every way. And they are absolutely useless unless they are being used to extend the abilities of human minds, and not just the minds of an elite group who can walk with them. Let them work.