跪拜 Guibai
← Back to the summary

HowToLiveBetter Puts a Price and an Evidence Grade on Every Life Decision

Open Source Project Issue #244: A High Cost-Performance Life Guide — 767 Pieces of Advice, Each with Cost, Benefit, Evidence Grade, and Original Source

Introduction

Today, we skip technology and talk about life.

This is the 244th entry in the “One Open Source Project a Day” series. Today's item is not a framework and not a tool, but a book: HowToLiveBetter. It lives on GitHub with 67,897 stars and 5,863 forks. The main text is published under CC BY 4.0, and the repository was created on September 7, 2026, less than two and a half months old as of today.

The book's subtitle is “Spend the least money, time, and energy to get back the most lifespan, money, and personal freedom.” It covers how to live longer, get sick less, and handle emergencies; how to avoid wasting money and which situations get people scammed or sued; what benefits are available when unemployed or broke, and what paperwork is needed to open a shop, start a company, or build a website; it also covers dating, marriage, children, going abroad, and learning a trade. In total there are 767 pieces of advice, each stating “what you spend, what you get back, and how solid the evidence is.” Sources cite only journal papers and official documents.

Why does a life-oriented repository belong in an “open source project” series? Because it applies “open source” to an unusual target: a knowledge base that can be searched, cited, corrected, translated, and queried by AI. It has issue-based errata, evidence grading, build scripts, and a skill that lets Claude Code and Codex answer questions according to the book.

What you will learn

Prerequisites

Project Background

Overview

This is a “menu of options, not a to-do list.” The README says this plainly at the start: you don't have to do everything; picking one or two items counts, and the author himself has not done most of them. It also warns that content may be outdated or wrong, that many policies are applied differently by locality, and that when you actually handle paperwork, see a doctor, or go to court, the local department, doctor, and lawyer are the authority.

This is what each item looks like (example from the README):

### 5. Switch the salt at home to low-sodium salt (potassium salt)
- Cost: a bag costs a few yuan more than regular salt. Swap it when you buy groceries; no extra time. Taste is almost unchanged.
- Plain language: people who have had a stroke or are over 60 with high blood pressure who switch to low-sodium salt have about a 12% lower chance of death within five years and about 14% lower chance of stroke.
- Benefit: a trial that randomly divided people into two groups, conducted in rural China with 20,995 people... the low-sodium salt group had about 12% lower mortality risk than the regular salt group (RR 0.88)...
- Evidence grade: A
- Source: Neal B et al. (2021). Effect of Salt Substitution on Cardiovascular Events and Death. NEJM. https://doi.org/10.1056/NEJMoa2105675
- Notes: disputed. The trial subjects were high-risk older adults; healthy younger people get much smaller benefits from switching salt. People with kidney problems... ask a doctor before switching.

Cost, plain-language conclusion, original benefit numbers, evidence grade, clickable source, and applicability boundaries are all present.

Author / Team

Project data

Quick Start

Most people do not need to deploy anything. The options below are ordered from least to most effort.

1. Use the online search page directly

Open the online search page. You can filter by keyword, chapter, or evidence grade, and also by three cost dimensions: whether it costs money, how much time it takes, and whether it requires willpower. Conditions can be combined. For the lowest-effort list, check the “extremely high” cost-effectiveness tier and get 123 items that cost no money, take little time, require no willpower, and land in the largest payoff bracket. There is also a life self-check: answer dozens of yes/no questions and see which items in the book you have not done yet, and which to do first.

2. Download an offline version

The repository's Releases offer four formats, automatically regenerated after every content update:

3. Let an AI assistant answer from the book

The repository includes a Chinese-language skill, skills/life-decision-guide, that supports Claude Code and Codex. After installing it, ask something like “Should I co-sign a guarantee for a friend?” or “Is a two-hour daily commute worth it?” It first looks up the relevant entries in the text, then sorts its answer using the book's cost-benefit method and notes which section and item each point comes from. If it cannot find something, it says so and does not invent numbers.

4. Run your own copy

git clone https://github.com/eternity4719/HowToLiveBetter.git
cd HowToLiveBetter
python -m http.server 8000
# open http://localhost:8000/ in a browser

The search page is pure static; README.md and book/ are its data. There is no backend, no database, and no dependencies to install. Note that index.html must be opened over http; double-clicking the local file will show a blank page, so use the offline single-file version in that case.

Core Analysis

Each piece of advice answers two questions: what you spend, what you get back

The book's most distinctive feature is that it writes life advice as a ledger. Every item must answer: what you spend (money / time / energy / willpower), and what you get back. What you get back is divided into four resources that are not convertible into one another:

The author states explicitly: mortality numbers, time-and-energy numbers, money numbers, and legal consequences are each counted separately and are not converted into one another. “Total mortality down 12%” and “save 500 yuan a year” are not on the same ruler and cannot be ranked. This restraint matters; it avoids a lot of false precision in the style of “save the cost of one coffee a year, compound it for decades.”

Who benefits is also graded: yourself, your spouse and immediate family, friends/colleagues/other relatives, and strangers. The first aid section should be read with this in mind: of 38,227 out-of-hospital cardiac arrests in China, 79.2% happen at home, so learning chest compressions is first about pressing on your own family. Rules like “don't jump in if you see someone drowning” and “don't get into a fight you see breaking out” are themselves self-protection rules. For whether to act to help a stranger, the book spells out liability exemptions, self-protection moves, and the risk surface, but does not do the math for you into a must-do obligation.

Evidence grades: A, B, C only answer “is the number trustworthy”

Grade Meaning Count
A A specific number can be checked; the source is a meta-analysis, a large cohort, or a randomized controlled trial, and it can state how much something dropped (HR, RR, or percentage decrease) 495
B Research supports it, but no exact figure can be stated; or only small samples or a single study exist 220
C The author's own experience, or a generally accepted practice, with no direct research literature 52

Another 88 items are marked “disputed,” and disputed A/B-grade items list the opposing evidence. All sources are primary: journal papers with DOI or PubMed links, or reports from official bodies such as WHO, CDC, and the National Bureau of Statistics; secondary retellings are not cited. For cases the government has not made public, a few items cite signed original journalism, with a web archive, and the notes say “media reports only.”

The author also adds a frequently overlooked reminder: A grade only means “a specific number exists and the source can be checked”; it does not mean that number is causal. The A grade includes both randomized trials and observational studies that only track people without assigning groups; the latter cannot distinguish whether the action worked or whether the people doing it were healthier to begin with. Which kind it is appears in the benefit field. For legal and policy items, A grade means the original statute or official document is cited.

How to read percentages is also explained: most are “a few tenths lower than people who did not do this,” not “a few fewer out of every hundred people.” With the same 12% reduction, older people who already have high risk lose far more deaths than younger people with low risk. So the instruction is: “first look at who is described in the plain-language line, then look at the percentage.” The end of the book includes a 41-item glossary (HR, RR, 95% CI, RCT, meta-analysis, confounding, reverse causation...), and on the online search page, hovering over words with dashed underlines shows an explanation.

Cost-effectiveness tier: the author admits this is a judgment, not evidence

Evidence grade does not answer “is it worth doing,” so each item also has a “benefit magnitude” (large/medium/small) and a “measurement category” (lifespan / money / time and energy / personal freedom), plus three cost dimensions. The search page combines these into a cost-effectiveness tier:

Tier Items Share
Extremely high 123 16%
High 344 45%
Average 300 39%

The magnitude has pre-set boundaries, not gut feeling: for lifespan, it looks at the percentage reduction (≥20% is large); for money, it looks at the amount (ten-thousand-yuan level is large); for personal freedom, it looks at consequences (avoiding criminal liability is large); for time and energy, it looks at what is saved (an hour a day saved is large).

The most honest sentence is: “This tier is the author's own judgment, not evidence,” and by the book's own standard it counts as only C-grade. It is separate from evidence grade. An item can be A-grade evidence and only average cost-effectiveness (the shingles vaccine has a phase III RCT showing 97.2% efficacy, but two doses cost three to four thousand yuan and shingles is rarely fatal), or C-grade evidence and extremely high cost-effectiveness (setting a lock-screen password on your phone and a PIN on your SIM card). The large number of items in the middle tier is also intentional: benefit magnitude only has three levels, and cutting it finer would be manufactured precision.

Engineering practices of a knowledge-type open source project

The engineering side of this repository deserves a separate look, because the maintainability of a “knowledge base” is often overlooked:

Project links and resources

Summary

This project does not teach a technique; it demonstrates how to write “life experience” as knowledge that can be checked: every item has a cost, a benefit, an evidence grade, a source, and a boundary; errors can be reported as Issues, and when updates arrive readers can get a new version in one click.

Three things are worth noting:

Evidence grading and cost-effectiveness tier are explicitly separated, and the latter is admitted to be only the author's judgment. Many “useful content collections” confuse “someone said this” with “evidence supports this.” This book uses A/B/C to answer “is the number trustworthy” and the cost-effectiveness tier to answer “is it worth doing,” and says plainly that the second one only counts as C-grade itself. That self-restraint is the main reason it is more credible than similar content.

The knowledge base is treated as software engineering. A single source of truth, automatic multi-format builds, public verification records, reader-submitted errata and translations, and a skill that lets AI assistants answer from the book with item citations: this set of practices is a useful reference for anyone who wants to turn documentation, manuals, or internal knowledge bases into living projects.

Read it with appropriate skepticism. This is a personal project written with the assistance of Claude Code. Laws, subsidy standards, and deadlines in the book can change at any time, and many policies are enforced differently by locality. The author himself reminds readers that when actually handling affairs, the local department, doctor, and lawyer are the authority. Also, I have not checked item by item whether all 767 numbers and conclusions in the repository match the original sources; I can only say that the author has provided checkable citations. Treat it as “a map for finding what to check,” not as the final answer.

If you want a clearly structured, source-traceable Chinese-language life reference, or you are thinking about how to turn a pile of knowledge into a searchable, reusable open source project, this repository is worth a look.

Explore PrimeSkills — curated AI agent and skill tools, each verified through real workflows. No hype, only genuinely useful tools.

Visit my personal homepage for more insights and interesting products.