> It rarely trends to maximization or the largest possible efficiency, since those conflict with "good enough".
Sometimes things get trapped in a local minima. Particularly when a seemingly inconsequential detail at a much much earlier stage becomes a dependency of lots of downstream stuff, but then it turns out that this just so happens to conflict with a better option in the here and now.
More commonly, the "perfect" solution is extremely brittle while the (supposedly) "good enough" solution is incredibly robust to all sorts of environmentally inflicted bullshit. In other words, most of the time evolution is practical while the humans criticizing the outcome are ignorant idealists.
I would go so far as to say that the vast majority of the time, systems that evolved are robust, not brittle, and you're right, this compromise "works better" or is "good enough to reproduce more than my relatives". And other times something gets caught in a local minima- but other bits around it optimize anyway (I think the "backwards" human eye might be an example of that- https://en.wikipedia.org/wiki/Evolution_of_the_eye#Placement and see also, https://en.wikipedia.org/wiki/Evolutionary_baggage).
Anyway, the example I was thinking of is here: https://en.wikipedia.org/wiki/Diffusion-limited_enzyme where some enzymes have evolved to reach extremely close to the maximum rate of catalysis limited by diffusion rates (and some enzymes have clever tricks to get around that).
Sometimes things get trapped in a local minima. Particularly when a seemingly inconsequential detail at a much much earlier stage becomes a dependency of lots of downstream stuff, but then it turns out that this just so happens to conflict with a better option in the here and now.
More commonly, the "perfect" solution is extremely brittle while the (supposedly) "good enough" solution is incredibly robust to all sorts of environmentally inflicted bullshit. In other words, most of the time evolution is practical while the humans criticizing the outcome are ignorant idealists.