Why I Keep a List of Problems That Won't Let Me Go

Kargi Chauhan · Aug 31, 2025 · 4 min read

Lately I've been thinking about why some people repeatedly do meaningful work while the rest of us feel busy but stuck. This is my current playbook — part confession, part reminder to myself on how I'm trying to aim my effort at the right things.

"Luck" is a multiplier, not a plan

Serendipity is real, but it sticks to people who are already moving. When I've felt "lucky," it was usually because I'd quietly done the unglamorous prep: built the toy prototype, read the dry appendix, kept a tiny dataset cleaned and ready. Then when the opening appeared, I wasn't starting from zero.

My operating rule: prepare as if luck will show up at 3 p.m. on a Tuesday — then it often does.

"Luck favors the prepared mind." — Richard Hamming

Work on important problems — ones you can actually attack

Not every "grand" problem is a good problem for me. There's a difference between inherently important and actionably solvable. I now ask two blunt questions before committing:

If this succeeds, would it move a needle I care about?
Do I have even a thin wedge of an approach today?

If either answer is no, I'm probably play-acting productivity. Important to me + a plausible first step beats grand-sounding + nowhere to start.

I keep a living list of 10–20 thorny questions

I maintain a private list of problems in my notes that won't let me go. Most are too big to "solve," but they're primed in my head. When a clue appears — in a paper, a podcast, a bug — I drop lesser tasks and chase it.

You can't be "first" on everything, but you can be ready when the trail marker finally shows up.

Courage is a daily micro-habit

The bold leap people celebrate usually rests on a stack of tiny acts of nerve: emailing someone out of my league, publishing a rough draft, admitting I don't understand a basic concept.

I've started rehearsing a simple line when I hesitate: do the brave small thing now. It compounds.

The "open door" advantage

Closed doors produce clean calendars and the wrong work. Every time I've sealed myself off to "focus," my ideas stagnate. The best ideas have come from letting myself get derailed by people who ask one inconvenient question.

So I optimize less for uninterrupted hours, more for collisions with curious minds.

Plant acorns, not monuments

Prestige is a trap: once you have some, it whispers, only do big, obvious things.

I try to ship small: a dataset with a good README, a tiny library with tests, a write-up that explains trade-offs. If my work requires me to babysit it, I've made myself indispensable in the worst way. The goal is usefulness without gatekeeping.

Style beats thrash

A drunk sailor's walk is how uncontrolled busyness feels: many steps, little distance. I block a weekly hour for nothing but big questions — no tools open, no tabs. I ask: Where is this field headed? What will likely be obsolete? What am I avoiding because it scares me?

That hour keeps me walking, mostly, in one direction.

Invert the problem

When I'm stuck, I flip the frame:

If the constraints were opposite, what would I do?
If I had to solve this with half the compute / data / time, what changes?
If the point wasn't just to "get an answer" but to produce a method others could reuse, how would I redesign it?

This habit has turned dead ends into paths more often than I'd like to admit.

Tolerate productive ambiguity

The trick is to be a believer and a skeptic at the same time: convinced the direction is right, unconvinced the current method is. That tension is uncomfortable — and it's where new approaches live.

If everything in my plan feels crisp and certain, I'm probably doing incremental work.

Read to find problems, not just solutions

Reading for answers makes me think like everyone else. Reading to spot gaps — missing baselines, brittle assumptions, untested edge cases — gives me original entry points.

I once watched a Waymo do a socially weird double lane change to avoid a tiny roadside object. Technically safe, human-unfriendly. That moment became a question I now carry: how do we evaluate "social predictability," not just safety constraints?

Presentation is part of the job

Good ideas don't automatically win. Clear writing, clean figures, and a tight five-minute spiel are force multipliers. I now budget at least as much time for the explanation as I did for the experiment.

If a smart outsider can't follow it, I haven't finished.

Change fields before the field changes you

When your instincts start auto-completing the next five years, it might be time to shift adjacency: new data regime, new evaluation, new interface, new collaborator. Not a complete reinvention — just enough to stop coasting.

I set a calendar reminder every year: "What would be uncomfortable to learn next?"

Is the grind worth it? For me, yes. The outcome matters, but the real return is who I have to become to do excellent work: clearer, braver, more generous, less precious. Direction over drift.

← Back to Kargi Chauhan