Deciding the Undecidable
In a recent blog post I introduced the notion of undecidable tasks — a particularly important type of work that’s not well covered by standard productivity advice.
These tasks are crucial to my job as an academic — as they are to many creative fields — so I devote a lot of attention to understanding how best to tackle them.
Today I want to share three tips along these lines that have worked well for me…
- Work on Undecidable Tasks Both During and Outside Work Hours. This type of task often requires a moment of inspiration where the pieces of a new approach suddenly click together. It’s useful, therefore, to not only dedicate regular workday deep work sessions toward their completion, but to also return to them, even if briefly, in unconventional settings more conducive to serendipity, such as while driving or walking the dog. I’ve found both types of thinking are necessary. A lot of intellectual progress can be made in structured sessions at the office while sometimes a hike in the woods is then needed to make use of this progress.
- Pursue (Exactly) Two Undecidable Tasks at a Time. Through hundreds of hours of experimentation I’ve found that having two undecidable tasks primed (see below) at a time is optimal. Two is better than one as it allows you to switch your focus if you get stuck (or fed up) with one task. But two is still small enough that your mind can keep the various pieces properly sorted and available for serendipitous reconfiguration.
- Undecidable Tasks Require a Decidable Priming. It’s not sufficient to have only a vague understanding of an undecidable task before you dive into solving it. You must first “prime” the problem by working out precisely: (a) what a solution would look like; (b) why standard or simple approaches fail; and (c) a sense of what type of approaches are promising and are worth exploring. This type of priming is a decidable task — something you can schedule and consistently complete in a fixed amount of time — and is something you must do before diving deeper on interesting problems. A non-primed undecidable task is merely a whiff of inspiration — not yet worthy of your limited time and attention.
The above tips have helped my work with the undecidable (i.e., my proof rate is higher when I structure my thinking in this manner). But simple heuristics are just scratching the surface of the fascinating — and under-discussed — intersection between undecidability and productivity.
Better understanding this type of work is something I plan to pursue in the New Year.
Great post Cal, I want to ask you if you are planning to write a book about what have you been writing the last past months or year, deep work. I think it would be a really interesting one.
Merry Christmas!
Pd: Im sorry for some spelling mistakes, English is not my mother language.
Stay tuned…
I was thinking the same! A book will be really helpful…
Cal , Recently i was having an interesting discussion on this topic with a friend. Slightly deviant from this but related topic is “Whether Computers would steal our jobs” . Watch the latest state of art on what computers could do, especially with advances in Machine Learning and Data Science. https://www.ted.com/talks/jeremy_howard_the_wonderful_and_terrifying_implications_of_computers_that_can_learn
The speaker touches upon this terrifying implications that people would lose jobs. So i quoted to my friend on your first blog post on this topic. My hunch is humans continue to have job due to undecidability stuff you are talking about
Read Tyler Cowen’s Average is Over for a nuanced view of this point. He (among others) argue that the rise of digital technologies will replace a lot of jobs, but will also make some (e.g., those for people able to work well with intelligent machines or market effectively) more lucrative. There’s definitely an undercurrent of undecidability in his discussion of what will still be valuable.
Hey Cal,
What you have mentioned in this post contradicts your earlier post “Work Less to Work Better: My Experiments with Shutdown Routines”. In that post, you mentioned that you followed a shutdown ritual that ensured that you did not think about hard problem after finish your work.But in the above post, you are encouraging to ponder on deep problems even during non-working hours. Isn’t this contradictory suggestion?
I think the shutdown routine post was a warning against going home and continuing to do the same work that we do during the day. Or alternatively, just worrying about that work. This post seems to be suggesting that hard problems need both structured thinking in the office and unstructured brainstorming in other locations. In other words, it’s not necessary to completely forget about our work once the day is over. But we shouldn’t just use the evening as extra hours in the daily schedule.
Great summary Duncan.
I was going to say the same thing.
Also I’m struggling to understand what undecidable tasks means exactly. My attitude towards productivity usually is: if process generates doubts about if it should be done, so this process is not so important in the first place. Because there is a difference between important tasks, essential tasks and extra tasks (which add value but can be neglected). In my view, undecidable tasks are not essential (because if they were, they wouldn’t be undecidable in the first place. Which means that they necessarily should be executed), maybe important or extra/bonus tasks.
Thanks!
I don’t think you quite have the right understanding of my use of the term “undecidable.” In my definition (from above), the term “undecidable” does not refer to the decision of whether or not the task is useful, it instead is a term from computability theory that references the fact that the process for completing the task is not something known in advance.
In other words, these are really important tasks that are not obvious how to tackle (e.g., solve a hard proof; write a beautiful chapter of a novel; come up with a turnaround plan for your company).
This is an important question.
To do decidable tasks (e.g., reply to e-mails) after your shutdown routine can reverse its positive effects. It is possible, however, to put aside time for undecidable task thinking outside of normal, shutdown book-ended work hours, and not lose the advantage of the shutdown.
I tend not to undecidable thinking after work on a normal 9 to 5 day (with the exception of the commute home, which I often use). But I do try to put aside time for undecidable tasks on vacations, weekends, and days off.
The key, I’ve found, is to choose in advance when to do this type of brainstorming work and keep it confined there.
For example, on a beach vacation, I might go for a long walk every morning to cogitate something tricky…but I know that’s it. This type of work, within these types of constraints, is not intrusive on your ability to relax and recharge, and can even be energizing.
Duncan already commented excellently, but I want to rephrase it still a bit.
I think unstructured thinking about deep problems happens automatically during out-of-office hours if (a) you have been thinking about the problem during office hours and (b) if your down-hours are not fully packed with other focus-consuming activities, but also consist of activities such as taking a walk, doing dishes, just sitting in the couch, etc.
I’m glad you mentioned “relax and recharge” because this is something that I think about a lot. We’re all busy and trying to achieve great success but its obvious that motivation and energy are finite. You can’t burn the candle at both ends forever.
I’m really interested understanding the process of regenerating energy and motivation.
I spend a lot of time trying to determine the best way to make a busy lifestyle sustainable however I haven’t come across many resources which tackle this. A lot of articles will simply say “Don’t forget to relax” or “Include time for rest” which doesn’t actually deepen my understanding of what is actually required by the mind/body to recharge.
So I was wondering whether anyone knew of a good discussion/treatment of the subject of recharging or energy levels? My own thoughts centre around adequate sleep (including naps where appropriate), regular exercise, regularly changing activities. But if you are worn out and exhausted (I have two kids under three), what’s the most “efficient” way to recharge your batteries?
I’ve been thinking about this kind of stuff a lot lately. Finding ways to categorize tasks so that I can properly schedule around them. Regarding step 2, the problem with planning two tasks is that then I start to hope that I’ll get both done, which distracts me and makes it more difficult to even get the first done. Instead I’ll schedule an hour or so for a single undecidable task, and have a list of more decidable “filler” task(s) lined up in case I finish up early.
Another thing that helps me is that I actually force myself to stop working when I hit the end of the time I’ve scheduled and move on to the next thing. It structures the day better for me so that I have to spend less time deciding around how and when to leave the project. I’ll schedule more time for it the next day if it seems important.
Nice post Cal.
We usually think of our work/non-work routines as a dichotomy verses as a trichotomy — work/non-work/reflective thinking. Reflective thinking generates creative solutions, the cross-pollination of ideas, and facilitates the “shutdown”.
Great posts, Cal. Happy New Year!
Cheers,
Mary
Hi Cal,
What I am fascinated by, is your ability to:
1. Be a world-class computer science theorist,
2. Actively blog (even on Christmas eve),
3. Read copiously,
WHILST raising a family (congratulations on the recent birth of Asa).
As a dad myself, working in the information industry (albeit from a programming side), and heavily involved in out-of-work activities, I am intrigued to know how you manage your time, specifically, how you spend quality time with your wife and children, amidst the myriad of activities you (effectively) swim in (did I mention – the ‘luxury’ of deep thinking?)
Paradoxically, the secret to my productivity is doing very little.
I confine my academic work to regular work hours, during which I maintain a deep commitment to deep work and minimizing the shallow.
Before Asa was born, I would blog once or twice a week after putting my older son to bed (after eight years of blogging and ten years of professional book writing, I can put out a post pretty fast). Now that Asa’s born, I’m doing to blog in little windows when I can get them…once he’s old enough to be put to bed reliably, I’ll return to my nighttime habit.
As for reading: I do a lot of it because I do almost nothing else. I don’t watch much TV. I don’t web surf. I don’t spend time at the gym (my workout happens in five minutes on a pull-up bar during my morning dog walk). In some sense, it’s my main extracurricular activity…so I get a fair amount done.
In other words, I’ve always found that doing less allows me to do more.