只能选两样:从来不及、会忘记到说不清——六个核心,看懂复杂世界的取舍 / Pick Two: Six Ideas for Understanding a Complicated World
Description
中文. 本著作是这一系列的入口读本:用六个核心,把六本学术专著写成一本从十岁到专家都能读完的书。
核心主张:看得见的世界背后有一张看不见的账单,账单只有三个价格——看得准、拼得上、传得下去,最多付得起两个。六个核心(看见、记住、跟上、选择、讲述、临界)就是学会读这六张账单。
结构与证据:19 部、123 页、29 幅图;四种读者分层(故事/多想一层/再想一层/给专家的注),每部配三个生活故事与动手练习,另有练习册、常见误解、给老师家长的一页与二十五个真实的人。
诚实的纪律:每个核心都标明对应专著的证据等级;凡通俗化损伤原意之处,以专著为准;本书不新增实证结论。
边界:它是入口,不是替代;具体结论与失败记录在六本专著中。它同时是系列的外部登记库(最小可验证单元、跨域盲测登记、合并术语表)。
版本:2026-10-18,123 页,19 部,29 图;源码开源,内容 CC BY 4.0,代码 MIT。
本版更新(2026-10-18 地基版)。本版把全书的共同前提写到卷首:现实中不存在无限——无限只作为极限出现在模型里,从不出现在现实系统的实际状态里。有限性本身只说明"有上界";它要变成"必须让出一样",还需要三个条件同时成立:需求随规模增长、描述/组合/传递共享同一承载者、同时维持两种活动更贵(交互惩罚 ψ>0)。因此"至多取二"一律是"紧预算(K/D 有界)下至多取二"的简写;若预算增长快于需求,三项可以同时满格。
本版新增内容。序章、六个核心一页速查、零基础读法与书末互引页的核心口号统一为"紧预算下,最多付得起两个";README 与《系列速读》新增"共同地基:有限性"一节,说明有限性是必要前提而非充分条件。
系列 / Series. 同一作者的独立研究系列:六本学术专著 + 两本通俗读本。入口与外部登记库:https://github.com/lyp0746/knowledge-tradeoffs;源码与可复算脚本:https://github.com/lyp0746/knowledge-tradeoffs。首选预览文件:knowledge-tradeoffs-print.pdf(印刷版)。
English. Abstract (English). This is the entry volume of the series: six core ideas turn six monographs into a book readable from age ten to specialist. Core thesis: behind the visible world lies an invisible bill with three prices — being precise, fitting together, and carrying over — and at most two can be paid; the six cores (seeing, remembering, keeping up, choosing, telling, criticality) teach how to read that bill. Structure and evidence: 19 parts, 123 pages, 29 figures; four reader levels (story / one layer deeper / one more / expert note), three short stories and a hands-on exercise per part, plus a workbook, common misconceptions, a page for teachers and parents, and twenty-five real people. Discipline: each core points to the evidence grade of its source monograph; where popularisation distorts the original, the monograph prevails; the book adds no new empirical claims. Boundaries: it is an entry point, not a substitute; it also serves as the series' external registry. Edition: 2026-10-18, 123 pages, 19 parts, 29 figures.
What is new in this edition (2026-10-18, Finitude Edition). The common premise is now stated at the front: there is no infinity in reality — infinity appears only as a limit in models, never as a property of a real system. Finitude alone only says that there is a ceiling; to yield a trade-off it also needs three conditions: demand grows with scale, description/composition/transfer share one bearer, and maintaining two activities at once is more expensive (interaction penalty ψ>0). "At most two" is therefore shorthand for "at most two under a tight budget (bounded K/D)"; if the budget grows faster than demand, all three can be full.
New in this edition. The opening, the six-core quick reference, the zero-formula reading and the series cross-reference now say "under a tight budget, at most two"; the README and the series reader add a section on the common foundation: finitude is necessary, not sufficient.
Notes
Files
knowledge-tradeoffs-ereader.pdf
Files
(10.4 MB)
| Name | Size | |
|---|---|---|
|
md5:7b932fa66305edd6002d18f8baaf9054
|
2.4 kB | Download |
|
md5:09cc5f00ff71ed38953d4944503e66d0
|
443.8 kB | Download |
|
md5:91a2dd2056703822777aa655a27b2cec
|
4.3 MB | Preview Download |
|
md5:69bafa29c130b2a04c25045c596e1165
|
397.5 kB | Preview Download |
|
md5:4b30795444bb7aa1f77f2da222ae5ce0
|
4.2 MB | Preview Download |
|
md5:5798bcbb5d547d23bc288764f2042df0
|
433.5 kB | Download |
|
md5:a84fac95cc3e7e05705a9974a2c93ea5
|
442.1 kB | Download |
|
md5:e7ec3b12b9f5fcec8b4aed8bf5deaedc
|
217.1 kB | Preview Download |
|
md5:eb6fae575de5542af3e270cb817a66f4
|
908 Bytes | Preview Download |
|
md5:22623a4417673c4a3c61d118a81195e6
|
7.1 kB | Preview Download |
|
md5:587644e6150596fb08dc87197070bfd4
|
1.2 kB | Preview Download |
|
md5:61218778b72fe2c1db875cb8c55fb0f1
|
281 Bytes | Preview Download |
|
md5:9b7de1f25d5b2e31917eab9d5d50cec2
|
204 Bytes | Preview Download |
Additional details
Subjects
- Complex Systems
- Popular Science
- Information Theory
- Cognitive Science
- Systems Science
- Scientific Methodology