How to decide what to work on today
Practical prioritizing with urgency and problem mapping
There’s a particular kind of stress that comes not from having too much to do, but from not knowing which of the many things matters most right now. I tried to find my way with systems—apps, frameworks, elaborate prioritization schemes—and eventually wrote software to calculate which tasks I should work on, using a formal tool from decision theory called expected utility.
The idea behind expected utility is simple. For any action you might take, you estimate the probability that it leads to an outcome, and you estimate how good that outcome would be (its “utility”). Multiply them together and you get the expected utility of that action—a number representing how much value you can expect to gain from doing it. Do this for every task, rank by the numbers, and work from the top.
I was proud of it. It felt rigorous.
Then I missed Thermodynamics Homework 4.
The assignment had been on my list all week. But its expected utility score was low—my software had no good way to represent urgency, so deadline pressure didn’t factor in. Meanwhile, tasks tied to bigger projects kept floating to the top. Homework 4 stayed buried until it was too late.
After that, I scrapped the software. And the failure forced me to ask a question I hadn’t taken seriously: What should I actually be optimizing for?
The answer I found wasn’t just a fix for the software. It changed how I think about every day.
The flaw in expected utility
The issue with my software wasn’t the math. Expected utility is a perfectly good framework. The issue was what I was feeding into it.
Assigning utility to tasks turns out to be extraordinarily difficult. How much utility is a homework assignment worth? How does that compare to an hour on a long-term project? I was essentially guessing—and my guesses reflected importance (big projects felt important) while completely ignoring urgency (deadlines didn’t register as “utility” the way they should). So the software optimized for importance and ignored time. Which is how a major assignment sat quietly at the bottom of my list while I worked on things that felt significant.
The lesson wasn’t “don’t use decision theory.” It was: you’re optimizing for the wrong variable.
What urgency actually means
That failure taught me how important urgency is. But besides a measure of a deadline, what is it?
Urgency is the degree to which time elapsing makes a problem unresolvable.
Why your productivity system keeps failing
After fifteen hours of work, I would watch an hour of TV before going to bed feeling like I’d wasted the entire day.
Think about what “missing Thermodynamics Homework 4” actually meant. Not just a missed assignment, but a worse grade in the class, which threatened my ability to pass, which threatened my bachelors degree, which I need for everything else I’m working toward. The longer I waited past the deadline, the more of those downstream aspirations got damaged. After a certain point, no amount of work could fix it.
That’s what urgency measures: how quickly options close.
This reframing matters because it connects urgency directly to your problem map. Every task you have serves some problem, some aspiration. An urgent task is one where delay causes irreversible damage to those aspirations. Not just inconvenience, but damage that can’t be undone.
Given that time is the resource I’m always short on, this makes urgency the right primary signal to follow. It’s not arbitrary stress. It’s the system telling you where time pressure is actively threatening what you care about.
There’s also a reasonable assumption baked in: urgent things tend to matter. A homework deadline damages my degree. A project deadline damages a relationship or a commitment. Urgency and importance aren’t perfectly correlated, but they’re correlated enough that urgency is a good first filter.
So the heuristic I use now is simple: if I think I can get everything done on time, I work on whatever is next. If I can’t get everything done on time, I focus on the most important among the urgent things, and accept that some less important things won’t happen.
Where urgency alone breaks down
But pure urgency has a failure mode that’s easy to miss until it’s too late.
Some of the most important things I’m working on have no natural deadline. Ganesha (the app I’m building) doesn’t have a ship date that anyone is holding me to. Learning to think and learn better has no due date. These things matter enormously for where I want to be in five years. But in a pure urgency system, they never rise to the top. Homework and commitments and agreements keep winning, week after week.
The issue isn’t that those urgent things are wrong to work on. It’s that a system driven entirely by urgency will let the important-but-not-urgent things drift indefinitely. Not catastrophically—it’s rarely “too late” in an absolute sense. But later than optimal, which compounds. A week of not working on Ganesha becomes a month. Aspirations that needed time to incubate—ideas that need to be slept on—don’t get that time.
The fix isn’t a separate “importance” metric that competes with urgency in some formula. Incorporating importance properly is the hard part—assigning it correctly, weighing it against urgency without double-counting, keeping it from swamping everything else—and getting it wrong is exactly what sank my software in the first place. The fix is something simpler: artificially creating urgency for important things before they drift too long.
What this looks like in practice: I have a recurring weekly task to work on Ganesha. Not because anyone is waiting on it, but because I decided it was important enough to protect with a regular commitment. Same with reviewing what I’m learning and how I’m learning it. The deadline is artificial—but it works the same way as a real one.
The deeper version of this is a regular review of your aspirations. Not just a quick glance at your task list, but actually going back to the problems you care about and asking: What matters most that I’m not currently working on? This isn’t a separate question you bolt onto your day. It’s the aspirations review I described in the last essay, now serving double duty—not just keeping the map current, but actively surfacing the things that urgency alone would leave buried.
I do this every morning. It takes a few minutes. But it’s what keeps Ganesha on my list alongside Thermodynamics.
What this looks like in practice
The full picture, then, is two things working together.
Day to day, follow urgency. Work on what’s due soonest, weighted by how much it matters. If you can do everything, do the next thing and enjoy the breathing room. If you can’t do everything, triage by importance among the urgent.
Regularly—ideally daily, at minimum weekly—review your aspirations top down. Not to second-guess every task, but to check: are the things that matter most actually getting time? If not, create artificial urgency. Give them a recurring slot, a self-imposed deadline, something that makes them show up in your urgency calculus.
This isn’t a perfect system. Expected utility, done correctly, would be better—it would let you make principled tradeoffs between urgent tasks and important-but-not-urgent ones, accounting for time and probability and the full downstream effects on your aspirations. That’s what I’m trying to build with Ganesha. But building it properly is hard, and I’m not done yet.
Until then, urgency is the best available heuristic. Not because it’s ideal, but because it’s the right variable to track given the constraint I’m always operating under: time. The urgent things are the places where time is actively threatening what I care about. Follow those first, protect the important ones with artificial deadlines, and review the whole picture regularly enough that nothing critical gets lost.
None of this works, though, without the problem map underneath it. Urgency without knowing your aspirations is stressful in a bad way because it feels that things are pressing without any sense of what they’re pressing toward. The aspirations review only means something if you’ve done the work of mapping what you care about and why. The heuristic is simple. What makes it useful is everything it sits on top of.
Written with the help of AI.