Church of Optimization 2: The Beverage
"Seattle is the best of both worlds: it has both a pretty big city and all the mountains and wilderness around." This is how someone justified to me, out loud, why they like living in Seattle. But wait, why are there a total of two worlds? And why am I supposed to care primarily about these two worlds? Just step outside of those two factors. Imagine someone cared about high fashion and thought that wearing your best flannels to a wedding is unacceptable. That person would not enjoy Seattle.
The statement about "the best of both worlds" conflates the ontology, it confuses a problem statement with an option of a solution, it confuses the question with the answer. "Where to live" is a problem each of us faces. "Seattle" is but one option for where you can live. Of course, in the value system of the person whom I quote at the top Seattle is the best place to live. However, I moved into Seattle, enjoyed it, and moved out of it for an entirely different set of reasons.
One friend of mine optimizes their life to support their motorcycle racing hobby, thus instrumentally making money for the garage workshop and the trips to world-famous racing tracks. Another one optimizes their life for the benefit of their dog, scheduling work and errands around canine daycare and entertainment. You can imagine that the two households look quite differently. Other people still try to achieve both the relative busyness of the city and the space of outer neighborhoods by moving to the suburbs and evangelizing The Suburban Way. Don't sell me suburbs. Suburbs are where life goes to die. I will not be taking questions on this point.
There is a common opinion that the supreme form of coffee is espresso. There is also a notion that there is a single correct, optimal way to do it, perhaps the way you would get in Italy. I had some espresso in Italy: it was pretty nice but I would not pretend my exploration was very thorough. In my last ten years in American academia, espresso machines gained a significant market share as a daily driver of science. I used to run into coffee enthusiasts who had a private espresso machine in their office and nerded out over cup preheating and puck preparation. By now many labs and break rooms have public espresso (not only Nespresso) machines with posted instructions; many grad students and postdocs know how to operate one. However, the perfectionist way of the singular espresso is not the extent of coffee. That is why the first time I saw a flight of three single origin espressos on a coffee shop menu (in Seattle!), I had it. I bought a bag of beans of one of those single origins to drink at home and found it to be much more enjoyable in a pourover. Paying the same money, I had a better experience by stepping out of the optimality criteria defined by someone else.
In the beverage industry, there are many options. Not only can you pick the perfect one for you, but take many at the same time. Whiskey flights, wine tasting, and 3oz cider glasses are celebrations of options. Many people balk at the prospect of ordering a glass of wine or a neat pour of whiskey at a bar since a bottle in the store is the same price. But trying a little bit of a beverage, or better comparing little bits of several options is how you learn. Stick with a familiar or try something new? In order to choose a bottle well, you need to do your homework on both the flavors and the prices. In formal modeling, this is known as the exploration-exploitation dilemma and shows up all over decision making science, from public policy to industrial design to reinforcement learning. This ode to options is catching so much popularity that I was recently offered a flight of mini-bagels paired with shmears!
Between the definitive bestness and the dazzling options, what can actually help you? One technique that is helpful is a coordinate scale. In Ann Arbor, MI, there is a pub called Ashley's which prides itself on offering nearly 100 beers on any day, with frequent rotations. The beer menu with complete listing takes many pages, but starts with an important preface on how to choose a beer. Even though I only went to Ashley's a handful of times, I learned the beer scale. While many novices evaluate beer options by their ABV, and advanced snobs by the IBU, a much better scale is something like "darkness". Darkness is the first principal component of the beer spectrum: the vast majority of beer varieties are somewhere between a weissbier and an Imperial stout (inclusive of bounds). A lager is near the beginning, a porter closer to the end, and a red ale pretty much dead center. Among these options, I don't have a persistent favorite but I know how to run away from the IPA-in-the-middle (IBU typically spikes mid-spectrum). On the rare days that I do feel like having a beer, I apply the Ashley's scale to the menu in front of me to choose something that fits the mood.
I argued before that optimization becomes a problem when applied to human affairs. It is even more of a problem when a person starts pushing their "optimal" solutions on others. If someone reached an optimal solution to a life problem suspiciously quickly, don't trust them: they do not have the depth of understanding, and probably not the breadth either. This reminds me of my early 20s when I wanted to write sweeping manifestos, to define the rules of life for me to live by, once and for all. I wish I were embarrassed to show those texts to the esteemed public today; however, there is nothing to show because unlike Lazarus Long I didn't have the true answers to life's problems and never finished writing my manifestos.
Moving past my 20s, I find that the pro-option attitude is generally more helpful than optimization. If you know many options, it is easier to make choices when your favorite is not available today. It is also easier to make recommendations to other people whose taste or skill set does not match your own. Solving the applied constraint satisfaction problem of choosing a conference dinner restaurant amid a group's varied dietary restrictions is not only responsible community service but also an intellectually fun challenge! One of the most useful academic writing guides I read did not preach one strategy of success but offered many that worked for different people. Premature optimization does more harm than good, and not only to you.