4 ms·
When I look at the things that I am most grateful to have experienced in my life, none of the decisions that would have been motivated by optimization processes
by titzer 9y ago
When I look at the things that I am most grateful to have experienced in my life, none of the decisions that would have been motivated by optimization processes would have led to them coming about.
So what, then, is the point of optimization criteria, if they wouldn't actually enable any possible decision making that accomplishes the "goal"?
Sorry to be so down on this, Kent, but this line of reasoning is such a fail on so many levels. Once in a while--maybe often!--we should just turn off the optimizer and let life unfold.
- tinymollusk 9y agoA standard explore/exploit model can work here. Perhaps it's semantics, but I consider that strategy an optimizer, although it's non a linear or deterministic one. Optimize the things I know are working; add semi-random decision making for some percent of my decisions to explore new spaces.