Why Should AI Give Me Life Advice? Turning a Book's Principles into Steps I Can Actually Take

There’s a book that’s been really popular lately, called The Cost-Effective Life Guide. Within a month of going up on GitHub, it had over 45,000 stars. What it cares about is very practical: living longer, not wasting time and energy, not spending money on the wrong things, and not getting into lawsuits. How to get sick less, what to do first when something goes wrong, which supplements and checkup packages aren’t worth buying, where to put money without getting ripped off, how a family gets through a major crisis—it covers all of it in detail, with dedicated chapters for elderly parents, kids in school, and running a business. There are more than 600 recommendations, each listing how much money it costs, how much time it takes, and what the benefit is, followed by research papers and official documents.

When I saw it, I knew this could be turned into a skill. But if I just loaded the book in and asked how I should live, it would still give me the answer for the average person. So I had Claude Code take the book, add what it knows about me, and give me advice. I use it for work every day, and it remembers my projects, my to-dos, and how my health is doing. The book doesn’t know who I am, but Claude Code does. Only by matching the two can it pick out the advice that genuinely applies to me.

But a lot of the advice depends on family circumstances. If something big happens at home, can we weather it? The book says that if savings can cover it, an emergency fund is enough as a backstop; only if they can’t is insurance needed. As my parents get older, what should I prepare ahead of time? I never talk about these things with Claude Code while working, and it can’t guess, so it had to ask me first. It gave me multiple-choice questions—what insurance I have, how old my parents are, how their health is—eight in total. Once I’d answered, it drew on the book to put together recommendations specifically for me, plus an extra section just about my parents.

At this point I still wasn’t satisfied. We already know all these truths, and still we can’t live this life well. So each recommendation had to be broken down into small things, bit by bit, and turned into a to-do list I could follow. For this step I added Tiny Habits (the Fogg Behavior Model) by Professor Fogg of Stanford, and blended the two books together: one tells me what to do, the other teaches me how to get it done. Fogg’s method comes down to three points: make it small, small enough to do even at my lowest; attach it to something I already do every day; and the moment it’s done, tell myself “Nailed it!” That little burst of good feeling stimulates the brain’s reward circuit and makes me want to do it again. It’s the same principle as getting hooked on short videos—only this time, it’s good habits I’m getting hooked on.

What to attach each habit to also comes down to Claude Code: the book doesn’t know how I spend my days, but it does. It paired every recommendation with a tiny habit. Three examples. I send Claude Code around 5,000 messages a month, and waiting for it to run tasks is the biggest stretch of downtime in my day—so every time I hit Enter to send it off on a long task, I stand up and walk to the window. It knows I eat a lot of takeout, so when I open the takeout App to order lunch, I swap the rice for mixed-grain rice. And at night, after brushing my teeth, I squat against the wall for ten seconds. Just ten seconds. This is a completely different thing from “exercise more”: when to do it, what to do, and when it counts as done are all crystal clear.

After this Life Checkup, I found quite a few blind spots—some I’d never thought of, some I’d thought about but not all the way through. Here are three. The first is family sharing of the medical-insurance personal account: once family members are linked, the money in the account can pay for my parents’ and kids’ doctor visits and medicine. I genuinely had no idea it could be used for family.

The second is getting vaccines and checkups lined up for my parents, who are in their seventies. The book recommends that older people get a flu shot every autumn, before flu season arrives in winter, at a few dozen to a little over RMB 100 a shot. For those 65 and over, it also recommends the pneumococcal vaccine, which some places offer free to seniors; after 50, there’s the shingles vaccine—two shots, blocking 97%, at RMB 3,000–4,000. Women 65 and over should get a bone density test, and if it turns something up, ask a doctor to prescribe medication and stay on it; some drugs now take only one infusion a year. None of this needs doing every day, so my to-do list only has the first step: next time I call my mom, ask whether the community flu vaccine has arrived, and book a bone density test while I’m at it.

The third is the legal risk of running an AI business. In recent years, China has introduced quite a few new requirements for AI businesses: providing generative AI services to the public requires a security assessment and algorithm filing first; generated images and videos must be labeled; and paid websites need the corresponding license. It listed these out one by one as a checklist, and I just go through them one at a time.

I’ve named this skill Life Checkup, and it’s open source. What’s truly useful about AI knowing us is this: it turns the principles in a book into the next thing we ourselves can actually do.

EP0115 · 2026-10-07 · 05:14. The picture uses a digital human, and the voice uses my cloned voice. This is an English translation of the approved Chinese script used to create the narration.

There’s a book that’s been really popular lately, called The Cost-Effective Life Guide. It’s only been on GitHub for a month, and it already has over 45,000 stars.

What this book talks about is really practical. It basically cares about a few things: how to live a little longer, how not to waste your time and energy, how not to spend money on the wrong things, and how not to accidentally land yourself in a lawsuit. And it goes into a lot of detail. Like how to get sick less, and what to do first when something goes wrong; which supplements and checkup packages you don’t need to buy at all; where to put your money so you don’t get ripped off; and if something big happens to your family, how to get through it. If you’ve got elderly parents, kids in school, or your own business, there are dedicated chapters for those too. There are more than 600 recommendations in total, and every single one says how much money it costs, how much time it takes, and what you get out of it. So is any of it backed up? Every one of them is followed by research papers and official documents.

When I saw it, I thought, something this good could totally be made into a skill—basically, having AI follow this book to help me get things done. But if I just loaded the book in and asked it how I should live, what it gave me would still be the answer for the average person, right?

So I told Claude Code: take this book, add what you know about me, and give me some real, concrete advice.

I use Claude Code for work every day, and it remembers the things I’ve talked to it about. My projects, my to-dos, even how my own health is doing—it knows all of it. The book doesn’t know who I am, but Claude Code does. Only by matching the two up can you pick out the advice that actually has something to do with me.

But quite a few recommendations in this book depend on your family situation. For example, if something big happens to your family, can you weather it? The book says if your savings can cover it, an emergency fund is enough as a backstop; only if they can’t do you need to buy insurance. So which one am I? Or, as my parents get older, what should I prepare in advance? To answer these, it first needs to know what coverage I have, how old my parents are, and how their health is. But these are things I don’t talk about with Claude Code while I’m working, so it doesn’t know, and it can’t just guess.

So it had to ask me first. It gave me multiple-choice questions—like what insurance I currently have, how old my parents are, how their health is. I answered eight questions in total. Once I was done, it went by what’s in the book and put together recommendations specifically for me, plus an extra section just about my parents.

But at this point, I still wasn’t quite satisfied. We already know all these truths, and still we can’t live this life well, right? So it has to actually be put into practice. That means breaking every recommendation down into small things, bit by bit, and turning them into a to-do list I can follow.

For this step, I added another book: Tiny Habits (the Fogg Behavior Model), written by Professor Fogg at Stanford, which is all about building small habits. So this skill is really me blending two books together: one tells me what to do, the other teaches me how to get it done. Fogg’s method really comes down to three points: make it small, small enough that you can do it even when you’re at your lowest; attach it to something you already do every day; and once it’s done, praise yourself right away: Nailed it! Why the praise? Because the good feeling right after you finish stimulates the reward circuit in your brain, so your brain remembers the action and wants to do it again next time. It’s actually the same principle as getting hooked on short videos—only this time, we’re getting hooked on good habits.

So what do you attach it to? Again, that’s where Claude Code comes in. Because what everyone does each day is different. The book doesn’t know how I spend my days, but Claude Code does.

So it designed a whole list of tiny habits for me, one for every recommendation. Let me give you three examples.

The first one is getting up and walking around. I send Claude Code around 5,000 messages a month, and waiting for it to run tasks is the biggest stretch of downtime in my day. So what it set for me is: every time I hit Enter and send it off on a long task, I stand up and walk over to the window.

The second one is swapping my staple food. It knows I eat a lot of takeout, so what it set for me is: every time I open the takeout App to order lunch, swap the rice for mixed-grain rice.

The third one is the wall squat. This one’s even smaller: at night, after brushing my teeth, squat against the wall for ten seconds. Just ten seconds.

Think about it—this is a completely different thing from just saying “exercise more.” When to do it, what to do, and when it counts as done—it’s all crystal clear.

So after doing this Life Checkup, I realized I had quite a few blind spots. Some I’d never even thought of; others I’d thought about, but not all the way through. Let me share three.

The first is called family sharing of the medical-insurance personal account. It means the money in your medical-insurance personal account, once you link your family members, can be used to pay for your parents’ or kids’ doctor visits and medicine. That this money could be used for family too—I honestly never thought of that before.

The second is getting my parents’ vaccines and checkups lined up. My parents are both in their seventies, and my mom’s heart isn’t in great shape either. The book recommends that older people and people with heart problems get a flu shot every autumn, ahead of flu season in winter—a shot only costs a few dozen to a little over RMB 100. And it’s not just the flu. For people 65 and over, it also recommends the pneumococcal vaccine, which some places give to seniors for free; after 50, you can also get the shingles vaccine—two shots, and it blocks 97%—it’s just a bit pricey, RMB 3,000–4,000. Then there’s osteoporosis: the book says women 65 and over should get a bone density test, and if it does turn something up, ask a doctor to prescribe medication and stick with it—some drugs now only take one infusion a year. None of this needs doing every day, so the to-do list only has the first step: next time I call my mom, ask whether the community’s flu vaccine has come in yet, and book a bone density test while I’m at it.

The third is the legal risk of running a business. I work in AI, and in the past few years China has introduced quite a few new requirements for AI businesses. For example, if you provide generative AI services to the public, you first need to do a security assessment and algorithm filing; generated images and videos need to carry a label; and paid websites also need the corresponding license. It listed all of these out for me one by one as a checklist, and I just have to go through and check them one at a time.

As for this skill, I’ve named it Life Checkup, and it’s already open source. The link is on my blog, xntj.tv. If you use Claude Code, you can go install it directly.

The way I see it, what’s truly useful about AI knowing us is this: it turns the principles in a book into the next thing we ourselves can actually do.

That’s it for today. See you tomorrow, bye.