Um I realized I haven't seen this argument written anywhere so here it is...

2nd law of thermodynamics says "in a closed system, entropy has to go up". Maximum entropy principle says "and moreover entropy goes up at the maximum allowable rate". This sounds like a seemingly minor difference, until you start thinking about what drives entropy up.

To make entropy go up faster, you have to have some kind of mechanism that drives it up. But mechanisms imply order! So you have to drive entropy down *inside* the matter of the mechanism in order to drive it up everywhere else.

Now if you really wanted to make entropy go up even faster, what would you do? Well you'd make a mechanism that not only drives entropy up, but can get bigger and bigger, probably by making copies of itself.

And whatever mechanism you started with probably isn't ideal, so you wouldn't just want to make exact copies, the way to drive entropy up as fast as possible is to run some kind of search for improved mechanisms that are even more efficient at driving up entropy, even more efficient at making more copies, even more efficient as searching for new mechanisms.

The good news is that the process of spewing entropy into the environment (which is the whole goal) is also exactly the process by which you can extract useful work from the environment. So the mechanisms have a really obvious source of fuel to drive themselves, running down the entropy gradient.

Now what does this kind of a thing sound like? Self-replicating mechanisms that maintain their own low-entropy state internally and learn by variance and selection -- basically the definition of life under evolution. And to do it even more effectively and efficiently requires intelligence, so it's actually intelligent life.

The 2nd law of thermodynamics is the "telos" of the universe, to the degree it has one. So somehow a slightly boosted version of the 2nd law of thermodynamics implies that the evolution intelligent life is part of that telos. Our "job" is to burn the gradient faster.

This can turn into a stupid e/acc worship-the-entropy-god thing but that's not actually what it implies. Evolution doesn't overfit like that. If you assume that there are always better and better ways to grow and live, that there is more intelligence and more possibility than you have imagined, then you have to ask "what is the guide?" Because the guide isn't burning as much fuel as you can today, it's learning the fundamental nature of the universe to be able to Dyson-cap stars, or build singularity engines, or whatever. And it turns out that the guide to that kind of growth is not single-mindedly mining at "get more energy" but rather ... interestingness. The thing that great scientists who make breakthroughs actually follow is their own curiosity. The great scientists themselves rely on instruments invented for all kinds of bizarre and seemingly unrelated reasons, no Newton without Kepler, no Kepler without Brahe, no Brahe without brass cannons and clocks and the Age of Exploration, no Age of Exploration without the Renaissance...it's all connected and the guide is interest.

Intelligent life is the machine the universe uses to blend itself, is one way to put it. But looked at another equally valid way the telos of the universe is to unfold along a path of maximum-interesting-ness. Because trying to maximize your self, your society, your world for maximum interestingness over time exports maximum entropy globally, they are the same thing.