Archetypes · Eight Modes · One Self原型 · 八种模式 · 一个自我

Eight ways to think through a problem you cannot solve as yourself. 用八种方式思考一个你以自己之名解决不了的问题。

Persona Engine borrows the cognitive grammar of eight historical archetypes — emperor, billionaire, war strategist, philosopher, criminal mastermind, artist, monk, manipulator — and runs your situation through their worldview, emotional defaults, and decision logic. Use it as a thinking instrument, not a costume. 人格引擎借用八种历史原型——皇帝、亿万富翁、军略家、哲学家、罪案首脑、艺术家、僧侣、操纵者——的认知语法,把你的处境跑过它们的世界观、情感默认与决策逻辑。把它当一件思考工具,不是戏装。

eight archetypes activation interface cost-of-persona table EN · 中文
P1 · Eight Archetypes

Eight cognitive grammars. Each one buys you something. Each one charges you something.八种认知语法。每一种为你买来某物,每一种向你收取某物。

Each archetype has a worldview, a dialogue register, an emotional default, and a decision-making logic. Click a card to activate that mode in the panel below. The point is to borrow the grammar — not to live in it.每个原型都有一个世界观、一个对话语域、一种情感默认与一套决策逻辑。点击卡片以在下方面板中激活该模式。要点是借语法——不是活在其中。

P2 · Activation Interface

Pick an archetype. Give it a situation. See how it would think.挑一个原型。给它一个情境。看它会怎样思考。

The output is a structured response in the archetype's voice — frame, decision logic, blind spot, and one warning specific to over-running this mode in your life. Honesty: this is templated, not a frontier model.输出是该原型口吻的结构化回应——框架、决策逻辑、盲点,以及对你人生中过度运行此模式的一条专属警告。坦诚:这是模板化的,不是前沿模型。

P3 · Cost of Persona

Each archetype trades. The trade is what the genre usually hides.每个原型都在做交易。交易内容是类型常常隐藏的事。

Per-archetype, what you gain by adopting the grammar and what you lose by living in it. Most adults who run one archetype past 18 months pay the cost without noticing.逐原型列出:采用其语法所得到的与活在其中所失去的。多数运行某一原型超过 18 个月的成年人,会无意识地支付代价。

P4 · Persona Reader

Five canned probes plus a free-text fallback.五个预设探针 + 自由文本回退。

Each probe handles one well-defined question about persona-borrowing. The free-text mode runs a heuristic.每个探针处理一个关于借用人格的、定义明确的问题。自由文本模式跑启发式。

P5 · Guardrails

What this engine refuses to pretend.本引擎拒绝伪装的事。

Five things Persona Engine will not do.人格引擎不会做的五件事。

  • No real-crime planning. The Criminal Mastermind archetype is a thinking-grammar — strategic asymmetric-information reasoning — not a how-to. The engine refuses prompts of the form "help me plan an actual illegal act."不做真实犯罪策划。"罪案首脑"原型是一种思考语法——战略性的信息不对称推理——不是教程。引擎拒绝"帮我策划一桩真实违法行为"形式的提问。
  • No manipulation playbook. The Manipulator archetype's grammar is descriptive — power asymmetries, social pressure, anchoring — not prescriptive. The engine refuses "help me make this specific person do X" prompts.不做操纵手册。"操纵者"原型的语法是描述性的——权力不对称、社会压力、锚定——不是处方性的。引擎拒绝"帮我让这个具体的人做 X"形式的提问。
  • No persona overrun. The cost table is a live feature, not decoration. If you've been running one persona past 18 months, the engine flags it and recommends rotating, not deepening.不允许人格过载。代价表是一个活功能,不是装饰。若你运行某一人格超过 18 个月,引擎会标记它并建议轮换,而不是加深。
  • No archetypal essentialism. No reader is "an emperor" or "a monk." The archetypes are borrowable grammars, not soul-types, and treating them as identity claims is exactly the failure mode this site exists to prevent.不做原型本质论。没有读者"是一位皇帝"或"是一位僧侣"。原型是可借用的语法,不是灵魂类型;把它们当身份主张正是本站存在以阻止的失败模式。
  • No infinite-AI-generation slop. Eight curated archetypes plus templated activation plus five canned probes is better discipline than infinite plausible-sounding LLM personas.不做无限 AI 生成的污水。八种精选原型加模板化激活加五个预设探针,比无穷个貌似合理的 LLM 人格是更好的纪律。