Greetings from Atlanta. đ
This weekâs newsletter elaborates on the trade-offs we face between exploration and exploitation, both in the small decisions we make day to day and over a lifetime. I also dispense with some basic teaching around open and closed questions. The essential takeaway: open questions create more expansive, dynamic, and rich conversations.
Iâve also made a few small changes to the newsletter. First, it will now be sent mid-week. Second, I plan to relax the newsletter template. It may continue to look like the old model most weeks, but my new intention is merely to include one big idea on a topic about questioning (or more generally, curiosity) and one more actionable item.
As always, thanks for reading. And if you havenât yet subscribed, you can do so here:
Exploration v. Exploitation
Computer scientists often speak of a trade-off between exploration and exploitation. The basic idea is that a machine learning algorithm will learn more if it explores many possibilities but it can often be more effective if it simply exploits the most promising one it discovers early on. The debate hinges on what search costs someone is willing to incur in pursuit of a more optimal solution vs. selecting a âgood enoughâ choice.
We confront a similar tradeoff everyday. Should we eat a favorite dish at our favorite restaurant or try the new place getting great reviews? Should we take the direct route home from work or a detour along unfamiliar (yet maybe more scenic) roads? Should we stay in our pretty good job or set out to look for a better one?
Mathematicians wrestling with these tradeoffs have developed a theory of optimal stopping. Iâll skip the gory proofs, but the key variable in making the âoptimalâ decision (as determined by either maximizing rewards or minimizing costs) is the interval over which we will enjoy the fruits of the action or choice. The closer we are to the end of a term, the more it makes sense for us to exploit the best option we have identified. To make this tangible, consider the choice of where to eat. The decision to try a new restaurant in our city might depend principally on how long we anticipate living in that neighborhood. If we can envision a moving date, we should probably make the rounds at all our favorite spots.
The math provides a simple heuristic for making a rational decision. Explore early. Exploit late.
But what if we stretch the interval to oneâs lifetime?
Our behavior certainly follows this arc. Children are explorers, and no parent would be surprised to learn that our questioning peaks around age 4 (right before we enter school). With each successive year out in the world, we gain a little more knowledge of how things work. By the time we settle into adulthood, we become ruthless exploiters. Exploration becomes either something we canât find time for amid our busy lives or have forgotten how to do. We see this not only in the habits we develop but in our worldview.
As Steve Jobs observed early in his career:
âPeople get stuck in these patterns, like grooves in a record.â
The theory of optimal stopping might suggest this to be appropriate behavior, and by and large, Iâd agree that exploitation should be the dominant tendency as we age. It is after all fundamental to our ability to contribute as productive adults that donât drift from curiosity to curiosity. But it is probably also fair to stay that we overshoot. We let the knowledge and mental habits we developed when we were younger guide us even as reality suggests we should revise our thinking. The world is too dynamic to let a stale mental model linger over an interval as long as a lifetime. But many of us let them. We need to be willing to explore, not merely to take advantage of an even better option we gave up looking for but to avoid a failing one. A general disposition to explore lets us update our habits or thinking.
Jobs famously borrowed the Whole Earth Catalogâs parting message when advising Stanford grads to âStay Hungry, Stay Foolishâ. In a sense, Jobs was recommending that students explore early AND explore late, even if it comes at the expense of the more direct, efficient route. We may sacrifice some traditional rewards like money and status. But an explorerâs life filled with questions can be pretty great too. We will likely learn and grow more. And we will probably have a lot more fun along the way.
What is your ideal balance of exploring and exploiting? Do you live that ideal now?
Closed and Open Questions
There are some elementary questioning concepts that Iâve delayed writing about since I launched this newsletter in May. I will start sprinkling them in over the next few months. When I do, I will try to find an angle that makes them feel a bit like a college level introduction rather than a middle school lecture.
The first topic is the difference between and role of closed and open questions.
Closed questions can be answered with a single word or short phrase. These include yes / no questions as well as queries like âHow old are you?â and âWhere do you live?â Closed questions tend to be easy to answer, quick to answer, and factual (setting aside common greetings like âHow are you?â). Sometimes closed questions are used to test for understanding or clarify, as in âSo youâre saying you prefer option A to option B?â
Closed questions begin with words like where, how many, who, when, or do you.
Open questions often require multiple sentences to answer. Open questions target knowledge, experiences, feelings or opinions. Our conversation partner is granted discretion over how to tackle our open question. This freedom generates unexpected answers, and more dynamic (and interesting) conversations.
Open questions begin with words like what, how, why, or tell me about a time.
There are a few additional dynamics to mentionâŠ
Open Questions are Expansive. They create dynamic, rich conversations with many branches for us to explore. If you want to have better conversations with family, friends, clients, or colleagues, consider how asking just a few more open questions can enhance dialogue. For example, rather than ask a partner âHow was your day?â, we can ask a variation of âWhat happened today that was unusual / memorable?â
Control of the Conversation. Closed questions allow the questioner to maintain some semblance of control of how a conversation unfolds. Imagine the lawyer methodically building his or her argument or the junior salesperson tunneling you toward a decision. Open questions create uncertainty about what might be said next. Even the most prepared individual needs to be learn to be comfortable responding in real-time.
Open and Closed in Tandem. Closed questions, especially those that elicit a ânoâ response, frame conversations. They telegraph what we want to talk about and establish guardrails that make our next question land more effectively. Consider last weekâs example âHave you given up on this project?â We can follow this question with an open-ended one like âWhat is standing in the way?â or âHow can we jumpstart it?â
Thatâs it for this weekâs dispatch. As always, please let me know what topics resonated with you or how I can improve the newsletter. If someone you know could benefit from this weekâs message, please donât hesitate to forward it.