To Break From Routine Is Human

Thursday, May 16, 2013
By Andy Brett

a penguin

This is crossposted on

The third James Bond movie, Goldfinger, opens with an amphibious mission to destroy an illicit chemical processing facility. After emerging from the water and planting the explosives, Bond strips off his drysuit to reveal a perfectly pressed white tuxedo and calmly affixes a red carnation to the lapel.

It’s rare that the next step after “take off your wetsuit” is “attach carnation to lapel” — it’s not routine, even ignoring the exploding chemicals bit. The Bond franchise and other filmmakers succeed by mastering this device, by leading viewers along a path and then suddenly yanking them out of the context they thought they were in. In some cases it’s overused to the point of becoming a trope (and ironically, somewhat predictable).

Another example is natural language processing — what Google uses to guess what you’re searching for, like it did just now when I typed “natural la”. In most sentences that start with “he left the keys on the kitchen…”, the next word is “table”, or maybe “counter”.

The drysuit-to-tuxedo reveal is the word that you don’t see coming. It’s the equivalent of completing the sentence above with “ceiling” instead of “table”. When you read a sentence or watch James Bond, your brain automatically produces guesses and predictions for what’s going to happen next.

When our guesses are wrong, we become engaged, or upset, or we laugh. We feel alive. That’s because bucking routine, and the expected, is uniquely human — for the moment, anyway. Artificial intelligence bots like Watson or competitors in the Turing test are usually confounded by the absurd or unexpected. It’s conceivable that they could eventually do quite well with it though, especially given how quickly humans turn the absurd into the hackneyed.

Avoiding routine

Falling into a routine can diminish your potential, not to mention your sense of free will. Athletes know that if they maintain constant mileage, weight, or intensity for a long period of time, their muscles will eventually stagnate. Much better is constant growth, followed by recovery.

How can you avoid falling into a routine? It’s helpful to have an outsider’s perspective. If you’re the one that’s stuck in the rut, it can be hard to notice and break out. The path of least resistance is usually the one that doesn’t bend much from the current trajectory.

There are plenty of hacks to get around this. The oldest, low-tech approach is regularly talking to a good friend who is willing to act as a sounding board and give candid feedback. A more formal approach looks something like a life coach — regular checkins on stated goals and aspirations.

In the life coach situation you’re paying, often dearly, for personal attention from a real human being in order to keep you on your toes. But what you get from a life coach (or any outsider, really) isn’t all that different from what you might tell yourself to do. If only you could have your present self communicate effectively with your future, soon-to-be-present self and guide him through evaluating some things in the cold light of yesterday.

Even before we get AI bots that can understand subtle, even British, humor, software is eating the world [1] in this domain, all the way at the top of Maslow’s pyramid.

That may seem like a ridiculous conclusion, and contrary to everything I just said — that you could write software to make people more spontaneous, more human, and more fulfilled. I suspect, though, that this falls into the category of problems where the right combination of AI and UI proves to be incredibly powerful.


Image: a variant of Tux the Linux Penguin


[1] I made a first attempt at an app that would add spontaneity to your day, with Whimsical — every day you get a new challenge to complete that will likely force you to break your normal routine. There’s also Everest, which is focused more specifically on goals that you set up for yourself and capturing “moments” as you progress towards them. Sort of a self-directed life coach. And of course there’s Beeminder, which is the ultimate way for your present self to get the attention of your soon-to-be-present self — through his wallet. Some kind of meta-beeminding goal could do the trick nicely.

Tags: , , , , ,

  • Ryan Freckleton

    This is a topic I’ve been thinking about recently, especially as it relates to beeminder. You’re exactly right that periodization (growth, then rest) and supercompensation (acute stress that then causes the system under stress to adapt) are necessary for human thriving. Virginia said a lot more about this with her Change Model.

    I hadn’t thought about having a meta-beeminding goal, but have debated whether I should start having goals that measure the standard deviation or kurtosis of activities I’m doing. One of the things I love about the “do more” style goals is that you can build up a buffer, so you can let your interests come in waves with beeminder simply putting a lower limit on what you’ll do.

    Another way to look at it that’s sometimes useful is you need separate phases of exploration versus exploitation. Routines are good at exploiting things, but exploration is a chaotic process.

    In our ancestral environment, a lot of this randomness and exploration came from the environment. A flood or drought would happen and you’d have to go find new places to hunt or plant crops. Our modern environment is a lot more chronic and it’s possible to get stuck in a rut. Boredom is one safety valve for this, boredom prevents us from doing anything too chronically. Using an app or something similar is a good way to inject randomness that’s outside of our control.

  • Andy Brett

    Great points Ryan. I really like the exploration vs. exploitation turn of phrase, it captures the idea well.

    I/we are still figuring out what a metabeeminding goal might look like. One approach would be to say “make me start a new goal every X days” but you’d wind up with a lot of goals that way and it seems easy to fudge. You could also a variant on the “must do” type of goal, where every [week|month|3 months] you enter a datapoint with a thought-out comment about whether you met your plan from the last time and what you’re going to do for this period.

  • Ryan Freckleton

    Another possible way to do it is with the “strech-churn” metric described by Cal Newport ( Basically, he measures the number of projects that he completes that put him out of his comfort zone.

    The issue I’m having is how to measure that period where you struggle with something new and that period of practice you need to integrate it into all of your other skills and behaviors, if you have one without the other, you stagnate.