Working on something important is often stressful. A hypothesis came up in conversation (h/t Michael Vassar) that labeling your work as important actually decreases the quality of the output. I have seen this effect in action quite a bit, e.g. when someone can write an essay when commenting on a blogpost, but gets nailed by writer’s block when writing an essay on purpose. Feeling responsible for something important can be an obstacle to productivity when the work requires creativity. How can we do important creative work without being held back by importance considerations?
This has come up in my thesis research project on variable selection. It’s been “almost done” for a year now, and my advisor keeps reminding me that it’s important to finish soon. This motivation results in making little tweaks to the algorithm at the expense of looking at the big picture. However, most of the improvements to the algorithm came about through exploring models that I thought were more interesting than the default one. When I came up with a more streamlined version of the model, for a while the algorithm was doing worse than the original one, and I started cursing myself for following my elegance heuristics instead of doing what needed to be done. Then I found a bug, and the new algorithm reached a similar performance level to the old one. (I do eventually want to stop playing with the model and actually publish the thing, though…)
I would like to distinguish between two types of importance: extrinsic and intrinsic. Extrinsic importance comes from external restrictions, like deadlines, funding availability, user demand or evaluation by high-status people. Intrinsic importance arises from the problem you’re solving – in my research example, this includes criteria like model simplicity and algorithm correctness. This is somewhat of a continuum, since a particular criterion can combine intrinsic and extrinsic importance, for example if my advisor insists on the algorithm being correct. Intrinsic importance fuels interest and curiosity, while extrinsic importance can inhibit creativity.
There is a quote by Howard Thurman: “Don’t ask yourself what the world needs; ask yourself what makes you come alive. And then go and do that. Because what the world needs is people who have come alive.” This can be interpreted as a suggestion to focus on intrinsic rather than extrinsic importance of your work. However, when choosing projects, it’s a good idea to be guided by extrinsic importance at least to some degree, lest you end up working on underwater basket weaving. Ideally, you’d want to find the intersection between the things that the world needs and the things that make you come alive. Throwing out the importance criterion once you start the project seems suboptimal, since it also applies to choosing subproblems within the project.
This can be helped by separating idea generation from idea evaluation, process focus from outcome focus. Write a sloppy draft of the essay, or imagine that you are writing to a close friend, and then step back and edit it based on the needs of your intended audience. Play with a new model for a while, occasionally asking yourself whether to prune this line of inquiry in favor of more promising ones. This process is analogous to a Kalman filter – an algorithm that alternates between an evolution step, where the next step is proposed, and an observation step, where that step is updated using data from the real world. In practice, much more time needs to be spent on developing the next step than on evaluating it, and more importantly, longer chunks of time. You want to correct your course once in a while, instead of stifling good ideas before fully developing them.
There is a saying that I find both annoying and insightful – “life is too important to be taken seriously”. Like many sayings, it’s annoyingly vague, but can be interpreted usefully. My interpretation of choice is that too much evaluation of what you’re doing can make the outcome less good, which conflicts with the purpose of evaluating it in the first place. “Life is too important to be taken seriously too often”, how about that?