Sublime

“knowledge frameworks.”

Any theory that proposes to reduce intelligence down to a single principle—or a single “master algorithm”—is bound to be barking up the wrong tree.

we will assume that agents will seek to maximize their average resource growth rate over many cycles of choice

Both are dynamic systems in which the selfish actions of countless individuals—whether they be cells or investors—lead to unpredictable consequences at the system level. In turn, these collective actions and consequences feed back to influence individual actions in endless cycles of adaptation and evolution.

contribute, and also the view that the collection of algorithms has not yet been well integrated; later in the book we will say why we are less optimistic about finding a single master algorithm.)

exploit the bugs

Second, there is the lack of robustness.

Look, sometimes you have to make quick decisions. You can’t just have humans make them; you have to combine human and machine methods.”