how to do great work, minus the cope

paul graham wrote a long essay called how to do great work. it’s good. you should read it. but you won’t actually do anything different after reading it, because the essay is built like a map and most people read it like a beach novel.

i’ve reread it maybe six times. each time i pull out one line and treat it like a task. that’s the only way it’s done anything for me. so this is less a critique of pg and more how i’ve started reading it.

the loop is the whole thing

pg’s argument, compressed in my head: pick something you’re genuinely curious about, get to the edge of what’s known, notice the gaps, fill them. repeat for decades.

that’s it. that’s the whole loop.

the part i kept missing on early rereads: step one (“genuinely curious”) and step two (“get to the edge”) have a 10-year onramp. it’s easy to bounce off in year two because the work doesn’t feel important yet.

my read now: it doesn’t feel important because it isn’t yet. importance is downstream of compounding, and compounding doesn’t show up on a timeline you can vibe-check.

curiosity is not a vibe

i used to read “work on what you’re curious about” and treat curiosity like an aesthetic. pick the topic that sounds good at a dinner party. it never worked.

the version that works for me: real curiosity has a tell. you keep doing the thing when no one’s watching, when there’s no reward, when you should be doing something else. you open the laptop on christmas break and feel better than you did all day.

if the thing i “love” requires a tutorial to start and a hype video to keep going, i don’t love it. i love the idea of it. for me those turned out to be very different things.

the question i actually ask myself: what do i do that doesn’t feel like work, that i’d do for free, that i’d defend in an argument with someone i respect? start there.

the edge is closer than you think

pg says: get to the frontier of your field. i used to picture a phd lab.

the way i think about it now: the edge is much closer than the word “frontier” makes it sound. for almost any niche, the edge is maybe 200 hours of focused work away. read the canonical books, ship five real projects, talk to ten practitioners. that puts you further along than most people who claim the title.

most “experts” i used to look up to weren’t at the edge of anything. they were at the edge of marketing. the gap between consuming content about a field and being at its frontier is smaller than the gap between thinking about starting and actually starting. the second gap is the one that ate years of mine.

notice the gaps

once you’re close to the edge, weird things start showing up. tools that should exist and don’t. workflows that everyone tolerates but no one fixes. questions everyone asks and no one answers.

my default used to be: someone smarter must already be on it. they’re usually not. the gap is yours because you saw it.

i’ve stopped waiting for “a great idea”. the move that actually works for me is to take the obvious gap i already noticed last week and spend a weekend filling it. that turned out to be the whole game.

work hard means something specific

pg’s “work hard” isn’t “grind 80 hours” the way the internet reads it. the way it lands for me: show up for the same hard problem every day until it cracks.

the people i watch doing great work aren’t doing more hours than me. they’re doing more consecutive days on the same problem. four hours a day on one project for two years beats 12-hour sprints between three abandoned ideas. not even close.

consistency beats intensity. the only way to compound is to not reset the counter. that’s the line i keep above my desk.

the prestige trap

pg buries this but it’s the part that rearranged my brain: do not pick your work for prestige.

prestige is a lagging indicator of work other people did decades ago. pick a field because it sounds impressive at parties and you’ve picked the field someone else made impressive. by the time you arrive, the interesting work is already done and you’re executing on someone else’s frontier.

the prestigious thing five years from now is something nobody respects today. that’s where i want to be.

the binary check

at the end of each year i ask one question:

is the work i did this year something i couldn’t have done last year?

if yes, the loop is working. if no, i spent the year in input mode dressed up as output.

great work isn’t a destination. it’s the side effect of running this loop for long enough that the compounding becomes visible to other people.

most won’t run it. the startup cost is real, the feedback loop is slow, and the prestige is delayed by a decade. that’s the whole reason it’s worth doing.

pick the thing. get close to the edge. fill the gap you already noticed. show up tomorrow.