"Play long-term games with long-term people." — Naval Ravikant
This week's premise: Long-termism is lonely because you're betting on a longer time horizon than the people who grade you. The company reviews you every six-month cycle, but the things that truly compound—deep skill, reputation, your network, reusable platforms—pay off two or three years out. So you keep choosing between "looks good this cycle" and "valuable in three years"—and when you choose the latter, nobody applauds. This week skips the "perseverance always wins" pep talk: long-termism has real costs—you can get overtaken in the short cycle, and you can sink-cost yourself into the wrong bet. Four tools: tell finite from infinite games and translate between them; spot what actually compounds; survive the valley of disappointment in the middle; and pick the right long-term people.
PRINCIPLE 01
Finite vs Infinite Games: First, Know Which One You're In
Finite vs Infinite Games
CarseTime HorizonTranslate
One-Line Principle
A six-month review is a finite game (it has an end, a winner, a ranking). A career is an infinite game (the goal isn't to win, it's to keep playing). Most of the pain comes from the wrong stance—playing an infinite game with finite-game intensity (burning the future for ranking), or playing a finite game with infinite-game patience (never shipping, so you're out this cycle). The real skill is knowing which game this is, and translating between the two.
Master's Words
"A finite game is played for the purpose of winning, an infinite game for the purpose of continuing the play."A career outlasts any single review; treat each as the game it actually is.— James Carse, Finite and Infinite Games, opening line
The Two Games, Side by Side
Scene
Situation: Halfway through the six-month perf cycle, your boss hints you "need a real highlight this period." You're holding a platform refactor that pays off in two years, and a flashy demo you could show off at review.
✗ Playing only the finite game
You go all-in on the demo and cut the platform investment. Review looks great, and you ride this playbook to a promotion—but two years later the tech debt detonates, the team fights fires daily, and nobody wants to work with you. You won a round and lost the match.
✓ Translate between the two clocks
"The platform refactor is a long bet; payoff lands at H+3."—and from that long line, carve out one milestone you can demo this cycle (migrate the most painful module first, show a measurable latency drop). Tell your boss plainly: which part is this cycle's delivery, which part is a bet on the future. Manage both clocks at once.
Checklist: Name the Game
Does this thing have a real "endpoint," or will it just keep going?
Is my current stance finite or infinite? Is it mismatched to the game?
Can this long bet be carved into a milestone visible this cycle?
Have I told my boss clearly: which part is delivery, which is a bet?
Common Mistakes
Treating everything as a finite game. Chasing only quantifiable short-term wins systematically burns future technical and relational capital.
Treating everything as an infinite game. Forever "laying foundations," never shipping—you're out this cycle, so there is no "later."
Not translating upward. Heads-down on the long line without explaining the payoff curve, and your boss just sees "no output this quarter." Long-termism still has to be seen.
Key References
James Carse, Finite and Infinite Games — the original frame. Simon Sinek, The Infinite Game — applied to organizations and leadership: "In an infinite game there is no 'winning,' only ahead and behind."
PRINCIPLE 02
The Shape of Career Compounding: What Actually Compounds
What Actually Compounds
CompoundingNavalTime Audit
One-Line Principle
Compounding isn't a platitude, it's math: early on you can barely see a difference, then after a threshold it pulls away exponentially. But not all busyness compounds—firefighting, repetitive work, and cleaning up after others are linear drains (done this week, reset to zero next week); deep skill, reputation, your network, and reusable artifacts are compounding assets. Pour your time out of the linear bucket and into the compounding one.
Master's Words
"All the returns in life, whether in wealth, relationships, or knowledge, come from compound interest."The asset you build quietly today is the leverage you cash in years later.— Naval Ravikant, The Almanack of Naval Ravikant
Scene
Situation: You work 50 hours a week. 35 go to meetings, firefighting, and cleaning up after others; 15 go to sharpening deep skills, writing reusable docs, and mentoring.
✗ Wearing "busy" as a badge
"I'm slammed every day, putting out fires everywhere—I'm essential."—Three years later you're in the same place: everything you were busy with was linear, reset to zero on completion, never accruing into an asset. The thrill of being busy masks a zero-compounding truth.
✓ Give every task a compounding check
Ask of each block: "Will this still be generating returns for me a year from now?" Firefighting → systematically reduce it (automate, delegate, hand the monkey back, say no to low-leverage asks). Mentoring / docs / deep skill → protect, even double down. Not work harder—pour the work into the bucket that compounds.
Checklist: Inventory Your Compounding Assets
Skill: Is there one capability that's clearly deeper and harder to replace than a year ago?
Reputation: Are there 3 people who'd vouch for me unprompted when I'm not in the room?
Relationships: Am I continually "depositing," not just "withdrawing" when I need something?
Artifacts: Have I left behind something reusable with my name on it (doc, framework, tool)?
Over the past month, roughly what was my compounding-to-linear time ratio?
Common Mistakes + A Note for Women
Mistaking linear busyness for compounding. "I grind" ≠ "I'm accruing"—the test is whether anything remains after you're done.
Compounding on the wrong thing. Going ever deeper on a dying tech stack is leverage on a depreciating asset.
Can't stand delayed returns. Compounding is flat early enough to make you doubt it—most quit before the threshold (see Card 3).
Female Leader's Note
Mentoring, cross-team lubrication, maintaining relationships—these are high-compounding assets, but they pay off slowly and are hard to quantify, so they often go uncredited in short cycles. Research and practice (Tanya Reilly's "Being Glue") show women are more likely to be defaulted into this "glue work / office housework": do it and it's taken for granted, skip it and you're "not a team player." The fix isn't to stop—it's to make the invisible labor visible: in your perf doc, write "the X mechanism I built sped up team delivery by Y," framed as leverage, not "I helped everyone."
PRINCIPLE 03
Holding Through the Doubt: The Valley of Disappointment
Surviving the Valley of Disappointment
James ClearThe DipCut vs Hold
One-Line Principle
The price of compounding is delayed returns: early effort and results are wildly out of proportion. You expect a straight line; reality is flat-then-steep—James Clear calls this stretch the "valley of disappointment." Most people quit one step before the threshold. But holding isn't blind: distinguish "not at the threshold yet" from "wrong bet, cut it"—white-knuckling the latter is just the sunk-cost fallacy.
Master's Words
"Habits often appear to make no difference until you cross a critical threshold... The most powerful outcomes of any compounding process are delayed."The steepest part of the curve sits just past the valley—right where most people give up.— James Clear, Atomic Habits, Ch.1
The Valley of Disappointment: Expected vs Actual
Scene
Situation: Six months into a platform migration, still no visible results. Stakeholders start asking "is this investment even worth it," and team morale is sliding.
✗ Two kinds of panic
Either you get spooked by the doubt and kill the project at the finish line (quitting before the threshold, wasting all of it); or you white-knuckle it but can't say why it deserves holding—just "we've put in so much, we can't stop now" (pure sunk cost).
✓ Replace "feelings" with "does the thesis still hold?"
"Does the core assumption we bet on still hold? If yes—look at leading indicators: did latency on migrated modules drop, did the failure rate fall? Prove with data that we're nearing the threshold, not chant 'just hold on a bit longer.' If the assumption no longer holds—cut decisively. That's not failure; it's redeploying resources to a better bet."
Checklist: A "Valley" or a "Dead End"?
Does the core assumption behind the original bet still hold today?
Are any leading indicators moving? (Even slowly—is the direction right?)
Do I want to hold because the thesis holds, or because "I've invested too much to let go"?
Starting from zero today, with no sunk cost, would I still launch this project?
If I drop it, can I redeploy the resources to a clearly better bet?
Common Mistakes
Sunk cost as a reason to hold. "We've put in 6 months" is not an argument—expected future return is.
Quitting one step before the threshold. The steepest part of compounding comes right after the valley—exiting at dawn is the costliest exit of all.
Treating every quit as failure. Seth Godin (The Dip): winners quit often too—they just quit fast and quit the right things. Strategic cutting is a skill, not surrender.
PRINCIPLE 04
Choosing Long-Term People: Treat Work as a Repeated Game
Play Long-Term Games With Long-Term People
Repeated GameAxelrodReputation
One-Line Principle
In a one-shot game, defecting/grabbing the advantage is the local optimum; in a repeated game, cooperation is. And with most colleagues you're actually in a repeated game—the industry is small, you'll cross paths again and again. Treat each interaction as a one-off transaction, and you win the short term while your reputation goes bankrupt long-term. The real leverage: repeatedly collaborate and deeply reciprocate with a few high-trust people.
Master's Words
"Play long-term games with long-term people. All returns in life come from compound interest — in relationships too."The PM you treat well today may be the director who hires you in three years.— Naval Ravikant, The Almanack of Naval Ravikant
Scene
Situation: On a cross-team project, the other side's PM did you a huge favor this time; but once it ends, you'll probably never work together again. And a peer you didn't get along with just left.
✗ Treating people as one-off resources
You decide they're "used up," and start ignoring their requests; or you quietly badmouth the peer who left. The industry is small: three years later that PM is the director of a team you want to join, and that peer decides whether you get in.
✓ Run reputation as a compounding asset
At a project's close, give feedback and credit publicly; even when you leave, leave gracefully (see Day 27 on the craft of exit). With a few long-term people, give first (Adam Grant's givers win long-term): help without demanding immediate return, building a "reservoir of trust" that cashes out only later.
Checklist: Inventory Your Long-Term People
Can I name 3-5 people I'll repeatedly work with and who are worth deep investment?
Over the past month, did I "deposit" more than "withdraw" with them?
Is there anyone I'm treating as a one-off resource whom I'll actually meet again?
When I leave a relationship/project, do I burn the bridge or leave room?
Common Mistakes
Treating all colleagues as one-shot games. The thrill of a small win trades for a bad reputation that slowly spreads across the industry.
Mistaking "long-term" for "infinite investment in everyone." Long-termism also cuts losses: a relationship that drains you one-directionally should be wound down.
Assuming long-term people never defect. They do. Use Axelrod's tit-for-tat: open with goodwill, respond clearly when crossed, but keep room to forgive—don't permanently blacklist after one defection, and don't get repeatedly exploited.
Key References
Robert Axelrod, The Evolution of Cooperation — "the shadow of the future": when reunion is likely, cooperation emerges on its own; tit-for-tat wins. Adam Grant, Give and Take — why givers win in the long run, and how not to get drained by takers.
This Week's Exercise · Your Day 31 Action
Do one concrete thing this week—not reflection, not reading:
Run a "compounding-check" time audit: log your work time blocks for a full week, tagging each as "linear bucket" (resets to zero when done) or "compounding bucket" (still generating returns a year out). At week's end, compute the ratio.
Then take just one action: pick your single biggest linear drain and find a way to cut it, automate it, or delegate it, and pour the freed time into something that compounds.
Reflection: If your review cycle changed from six months to once every five years, how would this week's priorities shift? That delta is exactly what short-termism is currently stealing from you.
Going Deeper
Can't "long-termism" become an excuse for present mediocrity and not shipping, justified by future returns?
Yes—this is its most common abuse. Telling real from fake long-termism turns on one thing: are there verifiable mid-course proofs. Real long-termism can produce leading indicators, carve out staged deliveries, and state a core assumption open to falsification; fake long-termism has only the slogan "it'll be valuable someday." A practical self-test: if you can't say "in three months, what specific signal tells me this path is right," you're probably not playing the long game—just using "long-term" as a fig leaf for procrastination. It doesn't exempt you from delivery; it just stretches the cadence—which still has to be seen and tested.
In a big company that reorgs and swaps your boss every ~18 months, are long-term bettors doomed to lose?
There's real tension, but it depends on where you anchor the long-term asset. Tie it to one specific project and a reorg can wipe you out; but skill, reputation, and relationships are portable—they travel with you and don't reset in a reorg. Strategy: in volatile environments, prioritize portable compounding assets and be wary of long investments locked to a single project or boss. Translating upward matters even more—when the boss changes, being able to quickly show with data "this long line is worth X and is at stage Y" beats being seen as someone with no recent output.
How do I tell whether I'm "crossing the valley of disappointment" or simply made the wrong bet and should cut?
Judging by feeling inevitably bends toward self-consolation; use two external anchors. First, does the core assumption still hold—did the market/tech change (it broke, cut it), or is it just not payoff time yet (unchanged, continue)? Second, are leading indicators moving—even slowly, is the direction right? A valley looks like "indicators improving but no visible inflection yet"; a dead end looks like "even leading indicators are worsening or flat." The most honest question: setting aside sunk cost, would I choose this path again from zero today?
Could deeply binding to a few "long-term people" actually factionalize me and get me labeled?
This is the real tension between Card 4 and org politics (Day 27): a tight reciprocal circle can look like a "faction" from outside. The resolution lies in binding to people, not camps—your long-term trust with someone rests on repeatedly delivered professional credit, not on "us against them." The former keeps you cross-cutting relationships through any factional reshuffle; the latter takes you down when the patron falls. In practice: long-term people can span different teams, connected by "good work and reliability" rather than allegiance; public credit and reciprocity (rather than private back-room deals) also let the relationship withstand sunlight and resist being labeled a clique.