Exploration/Exploitation Applied

In my previous post, I discussed the exploration/exploitation tradeoff that is inherent to optimizing under significant uncertainty, and alluded to some of the ways that this influences my thinking. In this post, I want to give some examples of how using this framework has actually changed my behavior.

From a Global Parameter…

As far as I can tell from using my introspection, it does seem like the exploration/exploitation tradeoff exists as something of a global variable in mind space. Where I am located along that spectrum changes the way I interact with the world. One notable example is how my curiosity functions. In exploration mode my curiosity is free floating, attaching to anything that catches my attention. This is Wikipedia plus tabbed browsing. In contrast, when I am in exploitation mode my curiosity is goal-directed. What specific skills do I need to accomplish the task at hand? I am very motivated to learn exactly what I need to know to make things happen.

My reward system even seems to function differently. In exploration mode, it feels intrinsically rewarding to learn new facts about the world, and even more rewarding to tie all of these facts together into some kind of coherent world model. In exploitation mode, what feels intrinsically rewarding is getting things done, of making a concrete plan broken down into substeps and methodically ticking it off line by line, of looking back at the end of the day and pointing to everything I have accomplished. It is useful to notice when I am in one mode or the other, so that I can utilize my respective strengths and figure out what is going to be fulfilling at the end of the day.

In my own case, I used to have this variable stuck all the way in “exploration mode”. This persisted (ironically) until my exploration led me to discovering this trade-off a few years ago, at which point I realized I was neglecting skill acquisition and concrete deliverables. This led to a period of rapid self-improvement, as I identified the lowest hanging fruit in my own life and optimized my situation.

…to Micro-Behaviors

On the opposite extreme, I notice this trade-off very acutely when looking at my habits of thought and behaviors. My tendency is for my behaviors to become ossified relatively quickly without active intervention – after an initial learning period (where no habits already exist), as soon as I satisfice my original criterion, I will settle into a routine and execute that pattern almost exactly time and time again. One example I like to use is how I shave. I will pretty much execute the exact same strokes in the exact same pattern. Is this method of shaving likely to be optimal? No. It takes me a while to shave right now, for example, and it’s possible I could cut down my time significantly by trying different techniques.

So how does thinking about this framework actually help me in practice? Because I realize when I have switched into exploitation mode, and I have some idea of how long I’ve been there. At that point, I can consciously choose to try something novel again, even if I’ve been doing something the same way for years. This gives me the opportunity to actually optimize the things I care about, instead of merely satisficing them.

And remember that this framework can be applied all the way up and all the way down. After each iteration of knocking myself out of a habitual behavior and trying something novel, I can update on my new solution. Did I achieve a substantially better solution? If so, further exploration is likely to be fruitful. Did I try doing this three times in a row without any results? Then perhaps I’m already close to an optimum. I will revisit habits multiple times over the years, but each time I do it I will let a little more time pass between trials. Over time I get more of a sense of when something is optimal or not, and how fruitful these sorts of explorations are, and adjusting my frequency of adjustment accordingly.

Mate Selection

One of the most important decisions we can get right is in picking our life partner. Relationship satisfaction is one of the strongest predictors of life happiness, as anyone who is in a difficult relationship will tell you. The majority of people do eventually settle down with someone (or more commonly these days a sequence of someones), so this is clearly a widely held human preference.

The bigger the potential payoff – and the higher the variance – the more time it is worth devoting to exploration. Given that this is one of the biggest decisions of our lives, you would think it deserves putting substantially more thought and time and energy into the search process! Once again though, the standard procedure is to execute by default: to spend time among the people you happen to come into contact with, and eventually mutually select someone in your vicinity. One of the most thoughtful approaches I’ve seen here comes from Spencer Greenberg, who echoes a lot of the advice I give to people.

Long story short, your strategy should be to try as many different things as possible as quickly as possible. Keep dating casual until you identify something of more potential, and potentially date more than one person at a time. Given that you may even have uncertainty about your preferences, it’s worth exploring things initially that don’t seem to be what you’re looking for. Even more importantly than knowing when to search is knowing when to stop – as soon as it is clear something will not work, it is best to end it immediately and try something new (Divia would probably say that it may be worth figuring out why it isn’t working so you don’t make that mistake again). Have a list of deal breakers ahead of time, and don’t settle quickly. These are all ways to increase the breadth of your search. Consider where you want to be at various points in life, and chain backwards form there: have kids by age X, so the marriage needs to be at age Y, so you want to be engaged by age Z, etc.

 Career Selection

If you followed the link about happiness predictors, you’ll know that job satisfaction is basically the other big predictor. A lot of the same considerations are true here as well as in love, though there are important differences. There is more bureaucracy involved in finding and leaving jobs, future employers tend to more rigorously consider your job history than potential partners do, work experience tends to be more domain-specific than relationship experience, etc. For all of these reasons, choosing what to do in life is in a lot of ways even more difficult.

As much as I am a generalist myself and rail against the seemingly increasingly strong trend towards specialization, the fact of the matter is that for most people specialization is going to provide the greatest return. It is entirely unclear that the rumor of Americans changing careers seven times in life is based in truth. So what does this mean? Exploration has to come much more up front in the job search. In fact, you might want to start considering different options before you even reach college. If you can get a jump start by doing internships or classes or related projects in your teenage years – or even earlier – you will be much more likely to find something you enjoy over the long term. You also look great to future employers, coming into an industry with more work experience than any of your peers.

It is also true that the world is facing accelerating change, so retaining some amount of flexibility is key. Being able to retrain if your chosen profession goes belly up will be increasingly important. Some skills are much more likely to be useful than others over the long haul too, so give that some thought before specializing in a declining industry.

It is worth noting that while all of this is true about choosing a career, it isn’t necessarily as true about choosing a specific job or employer. Switching companies too often does look suspicious, but ultimately your skill set will transfer relatively easily between similar jobs at different companies, and particularly early in your career it’s acceptable to still be searching for the right fit.

And I didn’t talk about entrepreneurship at all. That’s an entire series of posts in its own right…

The Framework Applied

In the above essay I took you all the way from the highest level mental framework down to the individual habits that compose our moment to moment actions, and how to think about two of the biggest decisions we will make in life. As you can see, I consider this trade-off to be widely applicable to life. One major benefit is in figuring out how much time I should spend considering options, relative to committing myself to a project. I personally am far too likely to consider a hundred options and never see any of them through, and thinking about this framework encourages me to pick something and stick with it. In practice though, it is much more common for me to employ this on the micro level, where I realize I am getting stuck in a habit and need to switch things up again.

Did you find this useful? Can you think of an area of your life to apply this to now? Post in the comments!