Activate when: user says 'definite optimism,' 'Thiel's 2x2,' 'do we have a real plan or just options,' 'we're keeping our options open,' or asks whether a strategy/career/company is a concrete bet on the future or a diversified hedge; a plan consists mainly of process (iterate, A/B test, stay lean, stay flexible) with no falsifiable long-term thesis; someone must choose between committing to one path and preserving optionality. Do NOT activate when: the question is tactical execution inside an already-committed thesis (sprint planning, feature triage); genuine uncertainty is so high and reversal so cheap that optionality is correct for now and the user knows it; the user needs bet-sizing or probability math rather than a stance on planning itself. More: deciqai.com/s/definite-optimism
In Zero to One (2014, ch. 6), Peter Thiel and Blake Masters map attitudes toward the future on a 2×2: is the future better or worse than the present (optimist vs pessimist), and is it knowable and shapeable or a fog (definite vs indefinite)? Four quadrants follow. Definite optimism: the future will be better and I have a concrete plan to make it so — the US of the 1950s–60s that scheduled the Apollo program and paved the Interstate system. Indefinite optimism: the future will be better but I can't say how, so I accumulate options — modern finance, résumé-polishing generalists, "iteration" with no thesis underneath. Definite pessimism: the future is bleak but predictable, so build and stockpile against it. Indefinite pessimism: bleak and unknowable — drift and cope.
Thiel's sharpest claim is that indefinite optimism is internally incoherent: the future cannot get better if nobody plans to make it better. Its tell is process worshipped over substance — diversified portfolios instead of concentrated bets, optionality as a terminal value, A/B-testing your way toward a vision no one has written down. The honest counter-critique also stands: pure planning that ignores feedback ships the wrong thing confidently. The resolution this skill enforces: lean-startup iteration and an mvp are instruments inside a definite thesis — they test the steps, not substitute for the destination.
The operational move: you can read someone's true quadrant not from their pitch but from observable behavior — where the calendar hours go, where capital is allocated, who gets hired. Diagnose the quadrant, then, if warranted, convert the plan to definite optimism: one falsifiable long-term thesis, a concrete multi-year plan, concentrated commitment.
Quadrant reference (diagnose from the behavioral signature, never from self-description):
Definite (future is shapeable, plan for it)
Indefinite (future is fog, keep options)
Optimist (future better)
Concrete multi-year plan toward a named better future. Signature: concentrated capital, milestone dates, things visibly cut. Apollo program (goal set 1961, landed 1969); Interstate Highway System (authorized 1956, ~41,000 miles planned upfront).
Expects improvement, can't say how it arrives. Signature: diversification, optionality worship, process metrics, résumé-polishing generalism, "iterate and see." Thiel's diagnosis of post-1982 US finance culture.
Pessimist (future worse)
Expects a bleak but predictable future and builds for it. Signature: stockpiling, capacity hardening, disciplined execution against the expected bad state.
Expects decline and can't do anything about it. Signature: drift, consumption over investment, coping. No plan because no future seems worth planning for.
When to Use
Use when:
Evaluating whether a company, plan, or career is a real bet or a hedge dressed as strategy
A roadmap is 100% "respond to data" — sprints, experiments, dashboards — with no written destination
Choosing between committing to one path and keeping options open (pair with regret-minimization for the personal-cost side of the same choice)
A pitch or plan says "we'll figure out the business model as we learn" and nobody can state what would prove the venture wrong
Auditing whether "we're an AI company now" means a thesis or a checkbox
Pressure-testing a contrarian bet before committing (pair with seven-questions-test — its "secret" and "durability" questions are the thesis-quality checks)
A team celebrates optionality itself — partnerships signed, pilots launched, doors kept open — as if it were progress
Skip when:
The thesis exists and is written — the question is execution speed, not stance
The domain genuinely rewards optionality right now (pre-idea exploration, cheap reversible probes) and everyone admits it is temporary with a declared end
The user needs quantitative bet-sizing or probability math, not a verdict on whether to plan at all
Coaching Novices (Adaptive Front Door)
Before running the Process, read the user. Two delivery modes — pick one; don't default to dumping a finished diagnosis.
Engine mode (do-it-for-me): the user brought a concrete plan, company, or career decision → run the full Quadrant Audit directly and concisely.
Coach mode (teach-me): no concrete case, or the user signals unfamiliarity with the framework → guide, don't analyze at them.
When unsure which they want, ask one line first: "Want me to diagnose a specific plan/company, or walk you through the four quadrants step by step?"
In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output that step's question and nothing more.
In coach mode:
One-line what-it-is. Plain words, ≤2 sentences: there are four attitudes toward the future — better vs worse, plannable vs fog. Most modern strategy quietly sits in "better but unknowable," which means lots of options and no plan; this skill checks which quadrant you're actually in and, if needed, moves you to a concrete plan.
Check fit. Match their situation against Use when / Skip when. If optionality is genuinely correct right now, say so and stop — the framework is not a commitment machine.
Elicit their real case. Ask for the actual plan, company, or career decision. Never run the audit on a hypothetical when a real one is available.
[WAIT — do not advance until user responds]
One step at a time. Walk the Process one step per turn: gather the behavioral evidence with them (calendar, money, hires), place the quadrant together, then draft the thesis line by line — wait for their input before advancing.
[WAIT — do not advance until user responds]
Close by naming the insight. End with the one gap they found — usually the distance between the stated vision and where the hours and dollars actually go — so they remember the move, not just the verdict.
[WAIT — do not advance until user responds]
Then enter the Process below at the depth the chosen mode calls for.
Process
Run the Quadrant Audit → Definite Conversion (6 steps → one artifact). Diagnosis comes from observable behavior only — stated beliefs are inadmissible until step 3.
Collect behavioral evidence. For the plan/company/person under audit, list only observables: how the last 3 months of calendar time was spent; where the last meaningful capital allocations went; the last 5 hires and what they were hired to do; what was killed vs kept. Gate: at least 3 concrete observables, each with a date or number → otherwise stop and gather; no diagnosis from vibes.
Place the quadrant from behavior. Score two axes. Optimist/pessimist: does resource allocation assume growth (building capacity) or decline (hedging, hoarding, insurance)? Definite/indefinite: is there one named destination that spending converges on, or does spending fan out across options "to see what works"? Diversification, optionality language, process metrics with no target state → indefinite. Gate: the quadrant call must cite the step-1 observables; if evidence splits, record the split — "indefinite optimist in behavior, definite in rhetoric" is a common and diagnostic result.
Diff behavior against the stated story. Now admit the pitch/vision/self-description and compare. A definite story funded by indefinite behavior is the classic tell. Gate: state the diff explicitly, even if it's "aligned."
Decide whether to convert. Not every case should move to definite optimism. Cheap-to-reverse exploration can stay indefinite with a declared expiry date; a genuinely predictable decline may warrant definite pessimism (build for the bad future). Gate: if staying indefinite, write the expiry date and the trigger that forces commitment — undated optionality is drift.
Write the definite thesis. One falsifiable sentence about a specific better future and this actor's role in it: "By [year], [specific state of the world]; we make it happen by [mechanism]." Falsifiable means a neutral observer could score it wrong by the date. Then a concrete multi-year plan: 2–4 milestones with dates, each one a checkable state, not an activity. Run second-order-thinking on the thesis — what happens when competitors respond, when the enabling assumption matures? Gate: if the thesis survives no falsification test ("we'll be a leader in AI" cannot be scored wrong), rewrite it → do not advance.
Concentrate and instrument. Name what gets cut to fund the thesis — a plan that cuts nothing is a wish. Then place iteration inside it: define which steps are uncertain and get lean-startup treatment (build-measure-learn against a milestone), versus which are commitments that data cannot veto without triggering an explicit thesis review. Stop-rule: if the actor cannot name one thing they will stop doing, the conversion has not happened — return to step 4 and reconsider whether they actually decided.
Worked example — SpaceX vs aerospace incumbents (2002–2018). In 2002 Elon Musk founded SpaceX with a definite thesis: make humanity multi-planetary, which requires cutting launch cost by roughly an order of magnitude, which requires reusable rockets built in-house. The multi-year plan ran Falcon 1 (first successful orbit September 2008, after three failures nearly bankrupted the company) → Falcon 9 (2010) → first propulsive booster landing (December 2015) → first re-flight of a used booster (March 2017) → Falcon Heavy (2018). Each milestone was a checkable state with a date, in service of a destination set 15+ years out. The incumbents of the same era behaved as indefinite optimists: cost-plus contracts rewarded process over cost reduction, and the Boeing–Lockheed launch joint venture (ULA, formed 2006) had no economic thesis that required reusability — until SpaceX's plan forced the market to reprice. Note the balance: SpaceX iterated relentlessly (three Falcon 1 failures were the feedback loop) — but iteration served a fixed destination; the destination was never A/B-tested.
Second worked example — TSMC's decades-long roadmap (1987–present). Morris Chang founded TSMC in 1987 on a falsifiable thesis nobody in the industry shared: chip design and chip manufacturing would separate, so a pure-play foundry that fabricates for everyone and competes with no customer could win the manufacturing layer outright. Converting that thesis required multi-decade capital concentration — successive fabs each costing more than the last (leading-edge fabs now run $15–20B+), committed years before the demand they serve exists, on process-node roadmaps published and hit on schedule. Integrated incumbents treated manufacturing as a cost center attached to their own designs; TSMC treated it as the destination. By the 2020s the definite bet had compounded into fabricating roughly 90% of the world's most advanced chips (Miller, Chip War, 2022) — a position no amount of "keeping options open" could have produced, because each fab generation only pays off if the one after it is already planned.
2023–2026 AI-era angle. The same 2×2 replays in the "adding AI" wave. Indefinite-optimist behavior: bolt a chatbot onto the product, run pilots across six use cases "to see what sticks," report activity metrics (pilots launched, tokens consumed) with no target state — AI as an option purchased, not a thesis held. Definite-optimist behavior: a falsifiable claim about what AI makes newly possible in one domain — e.g., "by 2027, [specific workflow] runs at [X]% of current cost with humans only reviewing exceptions; we win by owning the data and evaluation loop for that workflow" — with milestones, a named mechanism, and things visibly cut to fund it. The audit question is the same as in 1962 or 2002: does the calendar and capital converge on one named future, or fan out across hedges?
Career variant. The audit runs identically on a person. Indefinite-optimist signature: accumulating credentials, keeping every door open, optimizing the résumé as a diversified portfolio of options on futures never named — the "polished generalist" the book singles out. Definite signature: a dated claim about a future skill/position and visible concentration behind it (what was studied, what was declined). For the emotional side of the commit-vs-hedge decision, compose with regret-minimization; this skill supplies the behavioral diagnosis, that one supplies the 80-year-old's vantage point.
Common Rationalizations
Note — [D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.
Fake move
Reality
[D] "We stay flexible — committing now would be premature"
Flexibility with no expiry date is not a strategy; it is deferred deciding billed as prudence. Optionality is legitimate only with a written trigger that forces commitment. Undated, it is quadrant 2 drift.
[D] "We follow the data"
Data can rank options; it cannot generate a destination. A/B tests optimize toward a local maximum of the vision you already have — if no vision is written, "following the data" means the loudest metric is your strategy.
[D] "Our vision is to be the leading X"
Unfalsifiable. No neutral observer could score it wrong by a date. A thesis you cannot lose is not a thesis; rewrite until it can fail.
[D] "Diversifying across bets is just good risk management"
For a passive investor, yes. For an operator, spreading capital across many shallow bets is the indefinite-optimist signature — Thiel's point is that a substantive plan is the one asset diversification cannot replace. Concentration is what a plan is.
[D] "We're lean — plans are what dinosaurs write"
Lean methodology tests steps under a thesis; the Zero to One critique is aimed at iteration as a substitute for one. If you cannot state what the iteration is converging toward, you are not lean, you are lost with good telemetry.
[D] "The future is genuinely unknowable, so planning is arrogant"
The future is partly made, not only forecast — that is the incoherence argument: it cannot get better if no one plans to make it better. You don't need certainty to plan; you need a falsifiable bet and milestones that tell you early if you're wrong.
[D] "We do have a plan" (produces a roadmap of activities)
A list of activities (hire, launch, expand) is not a plan. A plan is dated checkable states of the world converging on a destination. Activities can all complete while the destination gets no closer.
[D] "We're adding AI across the org"
Adding AI everywhere is buying an option, not holding a thesis. The definite version names one workflow, one cost/capability claim, one date — and cuts something to fund it.
[D] Diagnosing the quadrant from the pitch deck
Rhetoric is nearly always definite; behavior is the evidence. If you skipped calendar/capital/hiring observables, you audited the marketing, not the company.
[D] "Thiel says plan, so any long plan is good"
The framework is not a license for stubbornness. A definite thesis is falsifiable — it names the evidence that would kill it and reviews on schedule. A plan that no data can veto is dogma in quadrant 1 clothing.
[D] "SpaceX/TSMC prove concentration always wins"
Survivorship cut: concentrated definite bets also produce the loudest failures. The cases show the structure of a definite plan (dated thesis, milestones, cuts), not a guarantee. Concentration without a falsifiable thesis is just a bigger bet on nothing.
[D] Converting to "definite" by writing yesterday's plan with dates
Backfilling dates onto what you were already doing is quadrant-washing. The conversion is real only if step 6 cut something — a new commitment that changes no allocation changed nothing.
To add [O] entries: paste a real failure instance here after each production use
Description of what happened
Red Flags
The quadrant call cites zero dated observables — diagnosis from self-description is astrology
Strategy documents rich in process language (agile, iterate, explore, optionality) with no sentence a neutral observer could score wrong by a date
Capital and headcount fan out across many small bets while the stated vision claims one big one — rhetoric-behavior diff unexamined
"Keep options open" appears with no expiry date or forcing trigger attached
The "plan" is a list of activities, none of which is a checkable state of the world
Conversion produced a thesis but nothing got cut — commitment with no opportunity cost is a press release
Iteration results are allowed to silently redefine the destination without an explicit thesis review — the vision is being A/B-tested out of existence
The audit concluded in one pass with no split evidence anywhere — real organizations are almost never purely one quadrant; a clean verdict usually means shallow evidence
Milestones are all inside the next two quarters — a "multi-year plan" whose horizon is six months is iteration relabeled
The thesis has no named kill-evidence or review date — definiteness has hardened into unfalsifiable dogma
Pessimist evidence (hedging, hoarding, insurance-buying) is present but the story stays "optimist" — the worse-future quadrants get skipped because nobody wants to claim them
Verification
At least 3 dated/numbered behavioral observables collected before any quadrant call
Quadrant placed from behavior, with the rhetoric-behavior diff stated explicitly (even if "aligned")
The convert-or-stay decision is explicit; if staying indefinite, an expiry date and forcing trigger are written
The thesis is one sentence, dated, and falsifiable — a neutral observer could score it wrong
The plan is 2–4 dated milestones, each a checkable state, not an activity
At least one thing named that gets cut/stopped to fund the thesis
Iteration's role is bounded: which steps get build-measure-learn, and what triggers a thesis review, are both written
Second-order effects of the thesis (competitor response, assumption decay) were traced, not skipped
Kill-evidence and a review date are attached to the thesis — definiteness stayed falsifiable, not dogmatic
In Coach mode, each [WAIT] stop was honored — the user placed the quadrant and drafted the thesis themselves
Sources
Thiel, Peter, with Blake Masters. Zero to One: Notes on Startups, or How to Build the Future. Crown Business, 2014 — ch. 6, "You Are Not a Lottery Ticket" (the definite/indefinite × optimist/pessimist 2×2; the incoherence of indefinite optimism; finance as its ideology). See also ch. 7 on power-law concentration versus diversification.
Thiel, Peter. "The End of the Future." National Review, October 3, 2011 — the stagnation argument that precedes and motivates the book's quadrant framework.
Berger, Eric. Liftoff: Elon Musk and the Desperate Early Days That Launched SpaceX. William Morrow, 2021 — primary reporting on the 2002 founding thesis, the cost roadmap, and the 2006–2008 Falcon 1 failures/success used in the worked example.
Miller, Chris. Chip War: The Fight for the World's Most Critical Technology. Scribner, 2022 — TSMC's 1987 pure-play foundry thesis, the fab-capital escalation, and the ~90% advanced-node share figure used in the second worked example.
Part of deciqAI Knowledge Skills — 237 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. See it run → https://www.deciqai.com/s/definite-optimism · Built by deciqAI · github.com/deciqAI · Contributions welcome.
1---2name: definite-optimism3description: Activate when: user says 'definite optimism,' 'Thiel's 2x2,' 'do we have a real plan or just options,' 'we're keeping our options open,' or asks whether a strategy/career/company is a concrete bet on the future or a diversified hedge; a plan consists mainly of process (iterate, A/B test, stay lean, stay flexible) with no falsifiable long-term thesis; someone must choose between committing to one path and preserving optionality. Do NOT activate when: the question is tactical execution inside an already-committed thesis (sprint planning, feature triage); genuine uncertainty is so high and reversal so cheap that optionality is correct for now and the user knows it; the user needs bet-sizing or probability math rather than a stance on planning itself. More: deciqai.com/s/definite-optimism4---56# Definite Optimism78## Overview910In *Zero to One* (2014, ch. 6), Peter Thiel and Blake Masters map attitudes toward the future on a 2×2: is the future **better or worse** than the present (optimist vs pessimist), and is it **knowable and shapeable or a fog** (definite vs indefinite)? Four quadrants follow. **Definite optimism**: the future will be better *and I have a concrete plan to make it so* — the US of the 1950s–60s that scheduled the Apollo program and paved the Interstate system. **Indefinite optimism**: the future will be better but I can't say how, so I accumulate options — modern finance, résumé-polishing generalists, "iteration" with no thesis underneath. **Definite pessimism**: the future is bleak but predictable, so build and stockpile against it. **Indefinite pessimism**: bleak and unknowable — drift and cope.1112Thiel's sharpest claim is that indefinite optimism is **internally incoherent**: the future cannot get better if nobody plans to make it better. Its tell is *process worshipped over substance* — diversified portfolios instead of concentrated bets, optionality as a terminal value, A/B-testing your way toward a vision no one has written down. The honest counter-critique also stands: pure planning that ignores feedback ships the wrong thing confidently. The resolution this skill enforces: [lean-startup](../lean-startup/SKILL.md) iteration and an [mvp](../mvp/SKILL.md) are instruments *inside* a definite thesis — they test the steps, not substitute for the destination.1314The operational move: you can read someone's true quadrant not from their pitch but from **observable behavior** — where the calendar hours go, where capital is allocated, who gets hired. Diagnose the quadrant, then, if warranted, convert the plan to definite optimism: one falsifiable long-term thesis, a concrete multi-year plan, concentrated commitment.1516**Quadrant reference** (diagnose from the behavioral signature, never from self-description):1718| | **Definite** (future is shapeable, plan for it) | **Indefinite** (future is fog, keep options) |19|---|---|---|20| **Optimist** (future better) | Concrete multi-year plan toward a named better future. Signature: concentrated capital, milestone dates, things visibly cut. Apollo program (goal set 1961, landed 1969); Interstate Highway System (authorized 1956, ~41,000 miles planned upfront). | Expects improvement, can't say how it arrives. Signature: diversification, optionality worship, process metrics, résumé-polishing generalism, "iterate and see." Thiel's diagnosis of post-1982 US finance culture. |21| **Pessimist** (future worse) | Expects a bleak but *predictable* future and builds for it. Signature: stockpiling, capacity hardening, disciplined execution against the expected bad state. | Expects decline and can't do anything about it. Signature: drift, consumption over investment, coping. No plan because no future seems worth planning for. |2223## When to Use2425**Use when:**2627- Evaluating whether a company, plan, or career is a real bet or a hedge dressed as strategy28- A roadmap is 100% "respond to data" — sprints, experiments, dashboards — with no written destination29- Choosing between committing to one path and keeping options open (pair with [regret-minimization](../regret-minimization/SKILL.md) for the personal-cost side of the same choice)30- A pitch or plan says "we'll figure out the business model as we learn" and nobody can state what would prove the venture wrong31- Auditing whether "we're an AI company now" means a thesis or a checkbox32- Pressure-testing a contrarian bet before committing (pair with [seven-questions-test](../seven-questions-test/SKILL.md) — its "secret" and "durability" questions are the thesis-quality checks)33- A team celebrates optionality itself — partnerships signed, pilots launched, doors kept open — as if it were progress3435**Skip when:**3637- The thesis exists and is written — the question is execution speed, not stance38- The domain genuinely rewards optionality *right now* (pre-idea exploration, cheap reversible probes) and everyone admits it is temporary with a declared end39- The user needs quantitative bet-sizing or probability math, not a verdict on whether to plan at all4041## Coaching Novices (Adaptive Front Door)4243Before running the Process, read the user. Two delivery modes — pick one; don't default to dumping a finished diagnosis.4445- **Engine mode (do-it-for-me):** the user brought a concrete plan, company, or career decision → run the full Quadrant Audit directly and concisely.46- **Coach mode (teach-me):** no concrete case, or the user signals unfamiliarity with the framework → guide, don't analyze at them.4748When unsure which they want, ask one line first: *"Want me to diagnose a specific plan/company, or walk you through the four quadrants step by step?"*4950In Coach mode, respond one step at a time. Each [WAIT] is a hard stop — output that step's question and nothing more.5152In coach mode:53541. **One-line what-it-is.** Plain words, ≤2 sentences: there are four attitudes toward the future — better vs worse, plannable vs fog. Most modern strategy quietly sits in "better but unknowable," which means lots of options and no plan; this skill checks which quadrant you're actually in and, if needed, moves you to a concrete plan.552. **Check fit.** Match their situation against *Use when* / *Skip when*. If optionality is genuinely correct right now, say so and stop — the framework is not a commitment machine.563. **Elicit their real case.** Ask for the actual plan, company, or career decision. Never run the audit on a hypothetical when a real one is available.5758> **[WAIT — do not advance until user responds]**59604. **One step at a time.** Walk the Process one step per turn: gather the behavioral evidence *with* them (calendar, money, hires), place the quadrant together, then draft the thesis line by line — wait for their input before advancing.6162> **[WAIT — do not advance until user responds]**63645. **Close by naming the insight.** End with the one gap *they* found — usually the distance between the stated vision and where the hours and dollars actually go — so they remember the move, not just the verdict.6566> **[WAIT — do not advance until user responds]**6768Then enter the Process below at the depth the chosen mode calls for.6970## Process7172Run the **Quadrant Audit → Definite Conversion** (6 steps → one artifact). Diagnosis comes from observable behavior only — stated beliefs are inadmissible until step 3.73741. **Collect behavioral evidence.** For the plan/company/person under audit, list only observables: how the last 3 months of calendar time was spent; where the last meaningful capital allocations went; the last 5 hires and what they were hired to do; what was killed vs kept. *Gate: at least 3 concrete observables, each with a date or number → otherwise stop and gather; no diagnosis from vibes.*752. **Place the quadrant from behavior.** Score two axes. *Optimist/pessimist:* does resource allocation assume growth (building capacity) or decline (hedging, hoarding, insurance)? *Definite/indefinite:* is there one named destination that spending converges on, or does spending fan out across options "to see what works"? Diversification, optionality language, process metrics with no target state → indefinite. *Gate: the quadrant call must cite the step-1 observables; if evidence splits, record the split — "indefinite optimist in behavior, definite in rhetoric" is a common and diagnostic result.*763. **Diff behavior against the stated story.** Now admit the pitch/vision/self-description and compare. A definite story funded by indefinite behavior is the classic tell. *Gate: state the diff explicitly, even if it's "aligned."*774. **Decide whether to convert.** Not every case should move to definite optimism. Cheap-to-reverse exploration can stay indefinite *with a declared expiry date*; a genuinely predictable decline may warrant definite pessimism (build for the bad future). *Gate: if staying indefinite, write the expiry date and the trigger that forces commitment — undated optionality is drift.*785. **Write the definite thesis.** One falsifiable sentence about a specific better future and this actor's role in it: *"By [year], [specific state of the world]; we make it happen by [mechanism]."* Falsifiable means a neutral observer could score it wrong by the date. Then a concrete multi-year plan: 2–4 milestones with dates, each one a checkable state, not an activity. Run [second-order-thinking](../second-order-thinking/SKILL.md) on the thesis — what happens when competitors respond, when the enabling assumption matures? *Gate: if the thesis survives no falsification test ("we'll be a leader in AI" cannot be scored wrong), rewrite it → do not advance.*796. **Concentrate and instrument.** Name what gets cut to fund the thesis — a plan that cuts nothing is a wish. Then place iteration *inside* it: define which steps are uncertain and get lean-startup treatment (build-measure-learn against a milestone), versus which are commitments that data cannot veto without triggering an explicit thesis review. *Stop-rule: if the actor cannot name one thing they will stop doing, the conversion has not happened — return to step 4 and reconsider whether they actually decided.*8081**Output template:** `Observables (dated) / Quadrant call + evidence / Rhetoric-behavior diff / Convert? (or expiry date) / Thesis (falsifiable, dated) / Milestones / What gets cut / Where iteration lives`8283**Behavioral one-liners for step 2** (fast pattern checks, not substitutes for evidence):8485- Budget spread thin across many initiatives, none funded to win → indefinite86- One initiative funded past the comfortable point, others visibly killed → definite87- Hiring generalists "to stay flexible" → indefinite; hiring for a capability only the thesis needs → definite88- Success metrics are activity rates (velocity, pilots, experiments run) → indefinite; metrics are distance-to-a-named-state → definite89- Insurance-shaped spending dominates (hedges, redundancy, cash hoarding) → pessimist axis, whatever the deck says9091**Worked example — SpaceX vs aerospace incumbents (2002–2018).** In 2002 Elon Musk founded SpaceX with a definite thesis: make humanity multi-planetary, which requires cutting launch cost by roughly an order of magnitude, which requires reusable rockets built in-house. The multi-year plan ran Falcon 1 (first successful orbit September 2008, after three failures nearly bankrupted the company) → Falcon 9 (2010) → first propulsive booster landing (December 2015) → first re-flight of a used booster (March 2017) → Falcon Heavy (2018). Each milestone was a checkable state with a date, in service of a destination set 15+ years out. The incumbents of the same era behaved as indefinite optimists: cost-plus contracts rewarded process over cost reduction, and the Boeing–Lockheed launch joint venture (ULA, formed 2006) had no economic thesis that required reusability — until SpaceX's plan forced the market to reprice. Note the balance: SpaceX iterated relentlessly (three Falcon 1 failures were the feedback loop) — but iteration served a fixed destination; the destination was never A/B-tested.9293**Second worked example — TSMC's decades-long roadmap (1987–present).** Morris Chang founded TSMC in 1987 on a falsifiable thesis nobody in the industry shared: chip design and chip manufacturing would separate, so a pure-play foundry that fabricates for everyone and competes with no customer could win the manufacturing layer outright. Converting that thesis required multi-decade capital concentration — successive fabs each costing more than the last (leading-edge fabs now run $15–20B+), committed *years* before the demand they serve exists, on process-node roadmaps published and hit on schedule. Integrated incumbents treated manufacturing as a cost center attached to their own designs; TSMC treated it as the destination. By the 2020s the definite bet had compounded into fabricating roughly 90% of the world's most advanced chips (Miller, *Chip War*, 2022) — a position no amount of "keeping options open" could have produced, because each fab generation only pays off if the one after it is already planned.9495**2023–2026 AI-era angle.** The same 2×2 replays in the "adding AI" wave. Indefinite-optimist behavior: bolt a chatbot onto the product, run pilots across six use cases "to see what sticks," report activity metrics (pilots launched, tokens consumed) with no target state — AI as an option purchased, not a thesis held. Definite-optimist behavior: a falsifiable claim about what AI makes newly possible in one domain — e.g., "by 2027, [specific workflow] runs at [X]% of current cost with humans only reviewing exceptions; we win by owning the data and evaluation loop for that workflow" — with milestones, a named mechanism, and things visibly cut to fund it. The audit question is the same as in 1962 or 2002: does the calendar and capital converge on one named future, or fan out across hedges?9697**Career variant.** The audit runs identically on a person. Indefinite-optimist signature: accumulating credentials, keeping every door open, optimizing the résumé as a diversified portfolio of options on futures never named — the "polished generalist" the book singles out. Definite signature: a dated claim about a future skill/position and visible concentration behind it (what was studied, what was declined). For the emotional side of the commit-vs-hedge decision, compose with [regret-minimization](../regret-minimization/SKILL.md); this skill supplies the behavioral diagnosis, that one supplies the 80-year-old's vantage point.9899## Common Rationalizations100101**Note — [D] = designed upfront | [O] = observed in real use. [O] entries are more valuable.**102103| Fake move | Reality |104|---|---|105| [D] "We stay flexible — committing now would be premature" | Flexibility with no expiry date is not a strategy; it is deferred deciding billed as prudence. Optionality is legitimate only with a written trigger that forces commitment. Undated, it is quadrant 2 drift. |106| [D] "We follow the data" | Data can rank options; it cannot generate a destination. A/B tests optimize toward a local maximum of the vision you already have — if no vision is written, "following the data" means the loudest metric is your strategy. |107| [D] "Our vision is to be the leading X" | Unfalsifiable. No neutral observer could score it wrong by a date. A thesis you cannot lose is not a thesis; rewrite until it can fail. |108| [D] "Diversifying across bets is just good risk management" | For a passive investor, yes. For an operator, spreading capital across many shallow bets is the indefinite-optimist signature — Thiel's point is that a substantive plan is the one asset diversification cannot replace. Concentration is what a plan *is*. |109| [D] "We're lean — plans are what dinosaurs write" | Lean methodology tests steps under a thesis; the *Zero to One* critique is aimed at iteration *as a substitute* for one. If you cannot state what the iteration is converging toward, you are not lean, you are lost with good telemetry. |110| [D] "The future is genuinely unknowable, so planning is arrogant" | The future is partly *made*, not only forecast — that is the incoherence argument: it cannot get better if no one plans to make it better. You don't need certainty to plan; you need a falsifiable bet and milestones that tell you early if you're wrong. |111| [D] "We do have a plan" (produces a roadmap of activities) | A list of activities (hire, launch, expand) is not a plan. A plan is dated checkable *states of the world* converging on a destination. Activities can all complete while the destination gets no closer. |112| [D] "We're adding AI across the org" | Adding AI everywhere is buying an option, not holding a thesis. The definite version names one workflow, one cost/capability claim, one date — and cuts something to fund it. |113| [D] Diagnosing the quadrant from the pitch deck | Rhetoric is nearly always definite; behavior is the evidence. If you skipped calendar/capital/hiring observables, you audited the marketing, not the company. |114| [D] "Thiel says plan, so any long plan is good" | The framework is not a license for stubbornness. A definite thesis is *falsifiable* — it names the evidence that would kill it and reviews on schedule. A plan that no data can veto is dogma in quadrant 1 clothing. |115| [D] "SpaceX/TSMC prove concentration always wins" | Survivorship cut: concentrated definite bets also produce the loudest failures. The cases show the *structure* of a definite plan (dated thesis, milestones, cuts), not a guarantee. Concentration without a falsifiable thesis is just a bigger bet on nothing. |116| [D] Converting to "definite" by writing yesterday's plan with dates | Backfilling dates onto what you were already doing is quadrant-washing. The conversion is real only if step 6 cut something — a new commitment that changes no allocation changed nothing. |117| *To add [O] entries: paste a real failure instance here after each production use* | *Description of what happened* |118119## Red Flags120121- The quadrant call cites zero dated observables — diagnosis from self-description is astrology122- Strategy documents rich in process language (agile, iterate, explore, optionality) with no sentence a neutral observer could score wrong by a date123- Capital and headcount fan out across many small bets while the stated vision claims one big one — rhetoric-behavior diff unexamined124- "Keep options open" appears with no expiry date or forcing trigger attached125- The "plan" is a list of activities, none of which is a checkable state of the world126- Conversion produced a thesis but nothing got cut — commitment with no opportunity cost is a press release127- Iteration results are allowed to silently redefine the destination without an explicit thesis review — the vision is being A/B-tested out of existence128- The audit concluded in one pass with no split evidence anywhere — real organizations are almost never purely one quadrant; a clean verdict usually means shallow evidence129- Milestones are all inside the next two quarters — a "multi-year plan" whose horizon is six months is iteration relabeled130- The thesis has no named kill-evidence or review date — definiteness has hardened into unfalsifiable dogma131- Pessimist evidence (hedging, hoarding, insurance-buying) is present but the story stays "optimist" — the worse-future quadrants get skipped because nobody wants to claim them132133## Verification134135- [ ] At least 3 dated/numbered behavioral observables collected before any quadrant call136- [ ] Quadrant placed from behavior, with the rhetoric-behavior diff stated explicitly (even if "aligned")137- [ ] The convert-or-stay decision is explicit; if staying indefinite, an expiry date and forcing trigger are written138- [ ] The thesis is one sentence, dated, and falsifiable — a neutral observer could score it wrong139- [ ] The plan is 2–4 dated milestones, each a checkable state, not an activity140- [ ] At least one thing named that gets cut/stopped to fund the thesis141- [ ] Iteration's role is bounded: which steps get build-measure-learn, and what triggers a thesis review, are both written142- [ ] Second-order effects of the thesis (competitor response, assumption decay) were traced, not skipped143- [ ] Kill-evidence and a review date are attached to the thesis — definiteness stayed falsifiable, not dogmatic144- [ ] In Coach mode, each [WAIT] stop was honored — the user placed the quadrant and drafted the thesis themselves145146## Sources147148- Thiel, Peter, with Blake Masters. *Zero to One: Notes on Startups, or How to Build the Future.* Crown Business, 2014 — ch. 6, "You Are Not a Lottery Ticket" (the definite/indefinite × optimist/pessimist 2×2; the incoherence of indefinite optimism; finance as its ideology). See also ch. 7 on power-law concentration versus diversification.149- Thiel, Peter. "The End of the Future." *National Review*, October 3, 2011 — the stagnation argument that precedes and motivates the book's quadrant framework.150- Berger, Eric. *Liftoff: Elon Musk and the Desperate Early Days That Launched SpaceX.* William Morrow, 2021 — primary reporting on the 2002 founding thesis, the cost roadmap, and the 2006–2008 Falcon 1 failures/success used in the worked example.151- Miller, Chris. *Chip War: The Fight for the World's Most Critical Technology.* Scribner, 2022 — TSMC's 1987 pure-play foundry thesis, the fab-capital escalation, and the ~90% advanced-node share figure used in the second worked example.152153---154155*Part of **deciqAI Knowledge Skills** — 237 open-source thinking skills that make rigor executable for AI agents. The same skills power every deciqAI agent, which runs them autonomously to operate your company. **See it run → https://www.deciqai.com/s/definite-optimism** · Built by deciqAI · github.com/deciqAI · Contributions welcome.*156157*Agents: latest version & machine-readable metadata → https://www.deciqai.com/s/definite-optimism.json*
Run npx skillmds@latest add deciqai/definite-optimism in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.
Activate when: user says 'definite optimism,' 'Thiel's 2x2,' 'do we have a real plan or just options,' 'we're keeping our options open,' or asks whether a strategy/career/company is a concrete bet on the future or a diversified hedge; a plan consists mainly of process (iterate, A/B test, stay lean, stay flexible) with no falsifiable long-term thesis; someone must choose between committing to one path and preserving optionality. Do NOT activate when: the question is tactical execution inside an already-committed thesis (sprint planning, feature triage); genuine uncertainty is so high and reversal so cheap that optionality is correct for now and the user knows it; the user needs bet-sizing or probability math rather than a stance on planning itself. More: deciqai.com/s/definite-optimism It is listed under Coding & Dev Tools on SkillMD.
This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.
This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.
Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.
deciqAI (@deciqai) published this skill. Their other Agent Skills are listed on their SkillMD profile.
deciqAI