用 AI 人的语言：关键期 ≈ 学习率调度。① 别把"软窗"当"硬窗"恐慌式排课——数学、编程没有"错过就没了"的窗口，硬塞反挤掉真正该趁早的事。② 真正的硬窗要认出来：若想孩子无口音地双语，语音敏感期确实有限，值得趁早；但据此推论"什么都要早"是范畴错误。③ 反推到你自己：成人大脑仍可重塑，"我太老学不会 X"对多数认知技能是自我施加的刹车，不是生物学事实——这恰是 AI 超级个体该破除的迷信。
English Summary
Critical & Sensitive Periods — A critical period is a hard window: certain neural circuits can be shaped by experience only within it; once it closes, the ability is nearly impossible to build. A sensitive period is a soft window: still learnable afterward, but at steeply higher cost. True hard windows are rare (mostly sensory and native-language phonology); most "skills" sit in broad soft windows or stay plastic for life. Crucially, window-closing is active, not passive decay — the brain applies molecular "brakes" to stabilize what's learned. This is the stability–plasticity dilemma, isomorphic to learning-rate annealing: a model that never anneals never converges; a brain that never closes never stabilizes a self. The real cost is misallocated anxiety — treating soft windows as hard ones.
AI Prompts
中文提示词
我在为 [孩子 / 我自己] 规划学习 [某项能力或技能]。请用"关键期 vs 敏感期"帮我判断：
① 这项能力到底是"硬窗"（错过几乎无法重建）还是"软窗 / 终身可塑"？给出依据；
② 如果是软窗，我现在的焦虑或排期是否用错了地方？哪些才是真正时间敏感的？
③ 给一个"学习率调度"式的建议：现在该重投入什么、什么可以从容来。
English Prompt
I'm planning learning of [ability/skill] for [my child / myself]. Use critical vs sensitive periods to help me judge:
1. Is this ability a "hard window" (nearly impossible to rebuild once missed) or a "soft window / lifelong-plastic"? Give your reasoning.
2. If it's a soft window, is my current anxiety or scheduling misallocated? What is actually time-sensitive?
3. Give a "learning-rate-schedule" style recommendation: what to invest in now vs what can wait.
Zone of Proximal Development & Scaffolding — The ZPD is the gap between what a learner can do alone and what they can do with guidance from a more capable other. Learning happens only in that gap: below it is boredom, above it is frustration. Scaffolding is temporary support across the gap, and its defining feature is fading: support that never withdraws is dependency, not scaffolding — a scaffold's goal is its own obsolescence. Isomorphic to curriculum learning and to the flow channel (challenge just above skill). Sharp corollary for the AI era: an AI that hands you the answer operates above your ZPD and atrophies your skill; an AI that hints, questions, and gives progressively less scaffolds within it. It's also a control problem — the scaffold must track the moving edge: help too much and you collapse the ZPD, too little and the learner falls out of it.
AI Prompts
中文提示词
我（或我的孩子）正在学 [任务 / 技能]，目前独自能做到 [现状]，卡在 [具体困难]。请把自己当成"搭鹰架的更有能力者"：
① 判断这个任务是否落在我的最近发展区内（太易 / 太难都要指出）；
② 用提示而非答案帮我跨越，分 2-3 个递减的支撑层级；
③ 给出明确的"撤除曲线"：每进展一步，你应该减少哪一类帮助。
English Prompt
I (or my child) am learning [task/skill]; I can currently do [status] alone and am stuck at [specific difficulty]. Act as a scaffolding "more capable other":
1. Judge whether this task sits within my ZPD (flag if too easy or too hard).
2. Help me cross it with hints, not answers, in 2–3 decreasing levels of support.
3. Give an explicit "fading curve": as I progress one step, which kind of help should you withdraw.
非平凡核心（悖论）：安全感反而促进探索，这与直觉相反——人们常担心"太给安慰会黏人"，事实正相反。确信基地一直在的孩子，探索得更大胆、走得更远。依赖与独立不是对立面：安全的依赖是自主的平台。神经层面：稳定的安全基地下调威胁系统（杏仁核 / 应激轴），把资源释放给探索系统——这是对"探索 vs 防御"权衡的一次神经门控。复杂系统视角：安全基地是个稳定吸引子，有它可退回，你才敢对状态空间做更大远征而不失稳。
① 育儿：对学龄孩子可靠的情感在场，会让他更独立、更敢冒险，而非更黏；至于"我刚发火了，是不是伤到他了"——别陷入完美主义，去修复，修复本身就是最好的一课。② 领导力：团队层面的同构是"心理安全感"——确信犯错会被修复而非惩罚的团队，探索与创新更多；做"安全基地"型领导，下属才敢冒险。③ 你自己 / AI 超级个体：为自己建一个"安全基地"——稳定的作息、身份与关系，正是这个吸引子让你敢在不确定中下大注。
English Summary
Secure Attachment — Early caregiver relationships internalize as a "working model" of whether others are reliable and responsive. A secure attachment makes the caregiver both a secure base for exploring and a safe haven to return to. The paradox: security enables exploration — a child certain the base is there explores more boldly, not less; secure dependence is the platform for autonomy. Neurally, a stable base down-regulates the threat system and frees the seeking system — a neural gating of explore-vs-defend. The least intuitive point: secure attachment is not constant presence or never letting a child be distressed; it's reliable repair ("rupture and repair"). Misattunement is normal; what builds resilience is the repeated experience that ruptures get mended. "Good enough" beats "perfect." Isomorphic to robust distributed systems: not never-failing, but reliably self-healing.
AI Prompts
中文提示词
背景：[描述一段关系——亲子 / 团队 / 自我]。最近发生了 [一次冲突或破裂]。请用"安全依恋"与"破裂-修复"框架：
① 诊断这是正常的失谐，还是真正的安全感受损；
② 给出一套具体的"修复"动作，而不是要求我做到完美无错；
③ 指出我可以如何成为更稳的"安全基地"，从而让对方（或我自己）更敢探索。
English Prompt
Context: [describe a relationship — parent-child / team / self]. Recently [a conflict / rupture] happened. Using secure attachment and "rupture & repair":
1. Diagnose whether this is normal misattunement or genuine erosion of security.
2. Give concrete "repair" actions — not a demand that I be flawless.
3. Show how I can become a steadier "secure base" so the other person (or I) dares to explore more.
① 育儿：别用奖励贿赂孩子做他本就喜欢的事；对他还不喜欢的事，用自主支持（讲理由 + 给选择）而非控制（贿赂 / 威胁）帮他内化。分清"结构"（好）与"控制"（坏）。② 你自己 / AI 超级个体：把热爱的项目变成纯粹的指标与变现追逐，可能挤掉当初让你出色的内在驱动——护好自己"探索系统"的自主性。③ 领导力：薪酬给足（移除不满因素），但别妄图用奖金"买"内在投入；给自主、给精进、给意义才是正道。
English Summary
Self-Determination Theory (Autonomy · Competence · Relatedness) — Intrinsic motivation rests on three basic needs: autonomy (volition, "I chose this"), competence (effectance, "I can grow"), and relatedness (belonging, "I matter"). Satisfy them → intrinsic motivation and well-being; thwart them → amotivation. The famous non-trivial finding is the overjustification effect: extrinsic rewards for an already intrinsically motivating activity crowd out intrinsic motivation — reward shifts the perceived locus of causality from internal to external. This is the human version of reward hacking in RL, and Goodhart's law in action: "when a measure becomes a target, it ceases to be a good measure." A crucial subtlety: autonomy-support ≠ permissiveness. Structure (clear expectations) supports competence; it's control (pressure, surveillance, conditional love) that kills motivation. Praise strategy/effort (informational), not fixed traits (evaluative). Relatedness — the often-forgotten leg — is the motivational face of secure attachment.
AI Prompts
中文提示词
我想激励 [孩子 / 团队 / 我自己] 去做 [某件事]。请用自决理论把我的激励设计审一遍：
① 它满足还是损害了自主、胜任、关系这三项？逐项打分；
② 我是否在对一件已有内在动机的事施加外部奖励（过度理由效应风险）？
③ 区分我用的是"结构"还是"控制"，给出一版"自主支持式"的替代方案（理由 + 选择 + 信息性反馈）。
English Prompt
I want to motivate [my child / my team / myself] to do [something]. Audit my incentive design with Self-Determination Theory:
1. Does it satisfy or thwart autonomy, competence, and relatedness? Score each.
2. Am I applying extrinsic rewards to something already intrinsically motivated (overjustification risk)?
3. Distinguish whether I'm using "structure" or "control," and give an autonomy-supportive alternative (rationale + choice + informational feedback).