Weekly Summary — 2026-08-10 to 2026-08-17

Overview

This week’s notes center on directing scarce time, attention, and judgment toward what compounds. They contrast surface-level activity—saving small amounts of money, collecting information, reacting to visible outcomes, or asking vague questions—with practices that change underlying capability: buying back time for growth, learning through application, investigating systems, and defining problems clearly. AI is framed as making prediction cheap rather than replacing judgment, which further raises the value of reliable context, decision-making, and execution. The health note grounds this broader theme in protecting sleep as a non-substitutable foundation.

Themes and insights

Scarcity is broader than money

Time, attention, trust, good opportunities, and favorable conditions are scarce alongside material resources. Several notes argue that small financial savings should be weighed against the time, energy, or attention required, and that freed time is most valuable when directed to learning, skills, rest, or relationships. Even abundance does not eliminate scarcity because desired quality, convenience, trust, and distinctive experiences remain limited.

Turn preservation and consumption into repeatable practice

The “bad apple” reflection distinguishes consumable resources from generative ones such as skills, relationships, attention, and love. Treating every valuable thing as inventory to save can defer lived experience; repeatable, adaptable practices can keep value active even as circumstances change. This complements the learning note’s emphasis on connecting, testing, and internalizing ideas rather than accumulating inputs.

Look beneath events to systems and assumptions

Visible outcomes are lagging signals, not necessarily explanations. The iceberg framework traces events through patterns and system structures to the mental models that create them; it recommends checking incentives, workflows, constraints, and leading indicators before treating symptoms. The same distinction applies to AI: prediction can generate a plausible signal, but judgment must account for asymmetric error costs, responsibility, and action.

Problem definition and informed inquiry are practical leverage

Good questions begin with independent research, clear goals and constraints, relevant evidence, and hypotheses or options to discuss. The preparation required often resolves much of the issue itself, while the remainder becomes easier for collaborators or AI to assess. This supports a learning loop of real problem → relevant concept → application → review → personal principle, rather than more passive consumption.

Protect the conditions that support judgment

Sleep is presented as a foundational condition for physical repair, emotional stability, cognition, and cardiometabolic health. The note emphasizes sufficient total sleep, regular timing, and a suitable environment rather than perfection.

Notes covered

• 0. Inbox/付费就是“捡便宜”】【付费就是“捡便宜” — Paying for delivery, transport, or expertise can be understood as purchasing others’ time and releasing one’s own. The note argues that the comparison is not simply price, but whether the reclaimed time can serve higher-value work, learning, rest, or family, with scalable knowledge products offering especially inexpensive access to accumulated expertise.
• “Ưu tiên Ăn quả táo hỏng trước phản ánh quan niệm sống nào?” — The “eat the bad apple first” dilemma cautions against spending disproportionate attention on preventing small visible waste or postponing valued experiences indefinitely. It proposes sustaining generative values through flexible, recurring practices rather than merely preserving possessions, while meeting present small losses without rehearsing future grief.
• Nguyên lý tảng băng chìm - nhìn kết quả, hiểu hệ thống — This note presents an iceberg model: mental models shape system structures, which create patterns that surface as events. Across business and other domains, it distinguishes lagging results from leading indicators and calls for investigating incentives, operating systems, economics, retention, cash flow, and other underlying causes before intervening.
• 学习不等于不断摄入信息 — Learning is defined as changes in understanding, judgment, and action, not ongoing information intake. The note proposes a loop of applying a relevant concept to a real problem, reviewing the result, and recording one’s own principle—particularly relevant when AI makes consumption and summarization effortless.
• AI is best understood as cheap prediction — AI is framed as cheaper prediction: filling missing information from known information, which also explains why plausible outputs may not be verified truths. As prediction-related tasks become cheaper, judgment, trustworthy data, and execution become more valuable; the note differentiates tool, decision, and strategy levels of adoption.
• 如何提出好问题 — Strong questions require prior research, a defined objective, context and constraints, and the asker’s hypotheses or options. The note treats questioning as collaborative problem definition rather than answer-seeking, preserves the legitimacy of asking for help after preparation, and applies the approach to technical work and AI use.
• 如何提出好问题 (2) — This companion learning-material version condenses the same case for questioning as a foundational ability and adds learning objectives, prompts, and flashcard candidates. It emphasizes preparation, discussion rather than extraction, and the need not to confuse improving questions with discouraging beginners from asking.
• Kinh Tế Học - Sự khan hiếm — Economics is grounded here in scarcity rather than assumptions that people are always rational or selfish. Scarcity includes intangible resources such as time, attention, trust, and favorable location, and persists as people seek better quality, safety, convenience, and status even when material production expands.
• 008 - 睡眠的底线与天花板 — Sleep is described as essential to restoration, emotional stability, cognition, and cardiovascular and metabolic health. The note recommends guarding adequate duration, keeping sleep and wake times regular, and maintaining a dark, quiet, clean, appropriately tempered environment.

Questions to carry forward

1. Which low-value demands on time or attention could reasonably be exchanged for resources, and what higher-value use would actually receive the freed time?
2. What knowledge currently remains at the input stage, and what real problem could test and turn it into a personal working principle?
3. For a recurring result or frustration, what pattern, system structure, incentive, or mental model may be invisible beneath the event?
4. When using AI-generated predictions or answers, what decision remains, who bears the cost of error, and what context or verification is needed?
5. What could be reformulated as a clearer question with a defined goal, constraints, evidence, and competing hypotheses?