I read this series some time ago, and I thought there were many gems in it that tracked my experiences that I could not articulate.<p>For instance, there's a bit in here about rationalism, and how the rationalist community might be good at being "correct", the kind of rationalism espoused, which prizes removing biases, doesn't always result in achieving the desired ends or winning.<p><a href="https://commoncog.com/blog/chinese-businessmen-superstition-doesnt-count/" rel="nofollow">https://commoncog.com/blog/chinese-businessmen-superstition-...</a><p>To me, a pure rationalism relies on your mental models of the world being correct in essential ways (to use a PCA analogy, in the principal components). However, the business world is one where higher-order effects dominate, so even if you have a rational first-order mental model of the world, that mental model is often insufficient.<p>This is where the empirical approach shines. The terms are used loosely in the article, but it differentiates between "epistemic rationality" (how do you know that your beliefs are true?) and "instrumental rationality" (how do you make better decisions to achieve your goals?), and the observation is that businessmen with high acumen -- in any culture, I would argue, but especially among the early Chinese settlers in Southeast Asia who generally lacked much formal education -- tended to employ "instrumental rationality" a lot. It was a kind of experimental culture within a conservative culture, which I find fascinating.<p>This seems to track with how startups work today -- create an MVP, ship, get feedback and iterate -- as opposed to the old product development approach of doing market surveys/focus groups, waterfalling the product engineering and then shipping a shiny product (that potentially no one wants).<p>I personally think that for most things, first-order mental models (i.e. simple cause and effect) will get you 80% of the way there, but in order to get over the lump of the last 20% of diminishing returns (where higher-order effects dominate), empiricism/instrumental rationality is necessary.<p>In the flip side, one of the downsides of being hyperrational is to overthink things. All mental models of the world are incomplete, so one can end up optimizing a narrow model of reality. We've all come across stuff that "shouldn't technically work" but do in real life. In many cases, it's more useful to just interact with reality directly. Uber and AirBnb shouldn't work -- some might argue they don't -- after all, who would trust strangers enough to book a car or a room on the Internet? Yet, empirically they do work for the most part, or at least the trust part does.