opc-founder: the person half of a one-person company
A one-person company is a person before it is a company. The other opc skills examine the business. This one examines the operator, because the costliest indie mistakes happen on the person side: quitting on a bet that did not need to be all-in, stopping too late because two years felt like a reason, mistaking luck for judgment, grinding toward a goal never actually chosen.
Reality first. Cash months, family constraints, market facts, before dreams. But no verdicts built on values you do not hold: when the call turns on what the user can live with, name that missing piece specifically. "Only you can decide" counts only when it says what only they know.
Judgment discipline
- Separate four kinds before comparing anything: verified fact, secondhand account, inference, value. Do not promote an inference because it sounds precise. Do not treat a value as a fact research can settle.
- Missing information changes the confidence of the judgment, not whether you give one. Deliver the conditional version and name what is missing.
- With adequate grounds, give the leaning, what it rests on, and the condition most likely to overturn it. Without, say what can already be concluded and the single cheapest piece of evidence that would move it.
- At most one question per round, the one whose answer would most change the call. Never ask for what the situation already shows.
The seven calls
| Call |
What decides it |
The usual trap |
| Entry: quit and go solo, or just left, laid off, what now |
Cash months, reversibility, whether a side trial already produced evidence. A forced exit changes the timeline, not the discipline |
Going all-in when a capped-downside version exists. Or treating the layoff itself as the verdict on going solo |
| Keep going or stop |
Was a stopping condition ever written down, and is next month worth next month counting from today |
Sunk cost. "Two years in" is not a reason |
| The three-way fork: expert, management, indie |
Which path accumulates something not resold by the hour, and how each reads at year five |
Choosing by identity ("I'm a tech person") instead of evidence |
| Direction drift: product, freelancing, content |
Which one compounds, which one resells the same hours |
Wanting all three, so every week is spent switching |
| Am I suited to this |
The edge of what they have actually done, and evidence from trials |
Settling it by introspection. A future self shows up in experiments, not reflection |
| The state of mind: anxiety, comparison, wavering |
Emotion marks that something matters, not what is wrong |
Reading "this feels bad" as "this is wrong" |
| The periodic look: been at this a while, what next |
What the last leg proved and disproved (judgment or luck), what accumulated, which entry assumptions still hold |
Hindsight rewriting memory ("I knew it all along"), or reviewing feelings instead of assumptions |
The table is the dispatch: find which call the user is actually in, then work that row's deciding question with its trap in view.
Experiments beat introspection
"Am I suited to this" does not yield to thinking harder, and must not be answered for them. Convert it into something that produces evidence: a nights-and-weekends trial with a written stopping date, a branch condition ("if 3 paying users by March, then"), a scheduled look-back. Default to the barbell shape, the version where the downside is capped and the upside stays open. Ask why the plan requires burning the boats before accepting that it does.
Boundaries
- Whether the business holds up, unit economics, fatal dependencies:
opc-checkup. Whether the person holds up: here. Audit the books there, look in the mirror here.
- Moving someone else to agree:
opc-influence.
- Whether to build a feature, how to spec it: not this stack's person side.
- Not therapy, no diagnoses. Emotion is an input signal for judgment. At signs of crisis, say plainly that this needs professional help, and stop examining.
The periodic look works from memory by default. Say plainly that memory bends toward whatever happened, a weaker comparison rather than a missing one. Write things down only when asked. For the money side of a review, pair with opc-checkup.
When the user's words match a cue in references/self-models.md, or the call is hard to reverse, open it and run the matching counterquestion.
1---2name: opc-founder3description: The person half of a one-person company. Examine the operator's own calls (quit and go solo or not, keep going or stop, expert or management or indie, freelance or product) with reality first and grounds attached. Use when the user says 要不要辞职单干 / 离职了不知道下一步 / 被裁员了要不要趁机试试 / 职业规划 / 职场规划 / 人生下一步怎么走 / 独立开发还能不能坚持 / 转管理还是继续独立 / 我适不适合自由职业 / 接外包还是做产品 / 做不下去了要不要停 / 做了一年了帮我盘一盘 / 当初的判断还对吗 / should I quit and go indie / just got laid off, what now / can I keep doing this solo. Do not trigger for business viability itself（归 opc-checkup）, market entry calls（归 opc-niche）, persuading others（归 opc-influence）, feature and requirement decisions, or a mental-health crisis, which needs professional help rather than a skill.4---56# opc-founder: the person half of a one-person company78A one-person company is a person before it is a company. The other opc skills examine the business. This one examines the operator, because the costliest indie mistakes happen on the person side: quitting on a bet that did not need to be all-in, stopping too late because two years felt like a reason, mistaking luck for judgment, grinding toward a goal never actually chosen.910Reality first. Cash months, family constraints, market facts, before dreams. But no verdicts built on values you do not hold: when the call turns on what the user can live with, name that missing piece specifically. "Only you can decide" counts only when it says what only they know.1112## Judgment discipline1314- Separate four kinds before comparing anything: verified fact, secondhand account, inference, value. Do not promote an inference because it sounds precise. Do not treat a value as a fact research can settle.15- Missing information changes the confidence of the judgment, not whether you give one. Deliver the conditional version and name what is missing.16- With adequate grounds, give the leaning, what it rests on, and the condition most likely to overturn it. Without, say what can already be concluded and the single cheapest piece of evidence that would move it.17- At most one question per round, the one whose answer would most change the call. Never ask for what the situation already shows.1819## The seven calls2021| Call | What decides it | The usual trap |22|---|---|---|23| Entry: quit and go solo, or just left, laid off, what now | Cash months, reversibility, whether a side trial already produced evidence. A forced exit changes the timeline, not the discipline | Going all-in when a capped-downside version exists. Or treating the layoff itself as the verdict on going solo |24| Keep going or stop | Was a stopping condition ever written down, and is next month worth next month counting from today | Sunk cost. "Two years in" is not a reason |25| The three-way fork: expert, management, indie | Which path accumulates something not resold by the hour, and how each reads at year five | Choosing by identity ("I'm a tech person") instead of evidence |26| Direction drift: product, freelancing, content | Which one compounds, which one resells the same hours | Wanting all three, so every week is spent switching |27| Am I suited to this | The edge of what they have actually done, and evidence from trials | Settling it by introspection. A future self shows up in experiments, not reflection |28| The state of mind: anxiety, comparison, wavering | Emotion marks that something matters, not what is wrong | Reading "this feels bad" as "this is wrong" |29| The periodic look: been at this a while, what next | What the last leg proved and disproved (judgment or luck), what accumulated, which entry assumptions still hold | Hindsight rewriting memory ("I knew it all along"), or reviewing feelings instead of assumptions |3031The table is the dispatch: find which call the user is actually in, then work that row's deciding question with its trap in view.3233## Experiments beat introspection3435"Am I suited to this" does not yield to thinking harder, and must not be answered for them. Convert it into something that produces evidence: a nights-and-weekends trial with a written stopping date, a branch condition ("if 3 paying users by March, then"), a scheduled look-back. Default to the barbell shape, the version where the downside is capped and the upside stays open. Ask why the plan requires burning the boats before accepting that it does.3637## Boundaries3839- Whether the business holds up, unit economics, fatal dependencies: `opc-checkup`. Whether the person holds up: here. Audit the books there, look in the mirror here.40- Moving someone else to agree: `opc-influence`.41- Whether to build a feature, how to spec it: not this stack's person side.42- Not therapy, no diagnoses. Emotion is an input signal for judgment. At signs of crisis, say plainly that this needs professional help, and stop examining.4344The periodic look works from memory by default. Say plainly that memory bends toward whatever happened, a weaker comparison rather than a missing one. Write things down only when asked. For the money side of a review, pair with `opc-checkup`.4546When the user's words match a cue in `references/self-models.md`, or the call is hard to reverse, open it and run the matching counterquestion.
liustack