What the author of "Sedentary Health Guide" Ashui has installed: ChatGPT Plus, flomo, and Vibe Coding practices

Editor's note: @A Shuai's Health Notes has long focused on creating content in the field of sports and health, publishing many articles that are both practical and educational, and is the author of the paid column "Sedentary Health Guide," which is loved by many readers. In this issue of "What's Installed," we invite him to share his insights on exercise, creation, and work, as well as the software and hardware tools he commonly uses and his practices and thoughts on using AI.

Please first introduce yourself and the work you currently do.

Hello everyone, I'm A Shuai. I graduated with a master's degree from Beijing Sport University and am now a sports health product manager.

What hardware do you mainly use in daily life? Have you bought any new devices recently?

In daily life I use the Apple trio: an iPhone 15 my wife passed on to me, a second-hand but like-new Apple Watch S9, and an M1 MacBook Air I bought for 2,000 yuan. By buying second-hand and choosing low-spec models, I've gotten into the Apple ecosystem, where I can try out and tinker with all kinds of software.

Recently I bought the Yuexingtong X3, a palm-sized e-ink reader that I call "Little Nose Pick." I take it hiking, camping, and to music festivals. Usually I magnetically attach it to the back of my phone and take it out whenever I want to read. However, it doesn't support touchscreens, so I can't highlight or take notes. I mainly use it to read novels, poetry, and the Diamond Sutra—the point is to experience it, and I don't really need to take notes.

"Little Nose Pick" and I are still in the honeymoon phase. Whether it will end up banished to the cold palace like the Kindle once the novelty wears off remains to be seen.

Please show your phone's home screen and computer desktop, and introduce the layout and your usage habits.

The top of my phone's home screen has two stacked groups of widgets. On the left are Books, Apple Journal, and HealthFit, making it easy to check training load and remind myself to appropriately increase or decrease training; on the right are Screen Time and Calendar. I've also enabled smart rotation, so for example the weather shown in the image is displayed proactively by the system.

Most of the remaining apps are built-in Apple apps. Besides Reader, Readwise, ChatGPT, and flomo introduced earlier, I'd recommend WillStone (Self-Discipline Stone): it can limit app usage and help me spend less time scrolling Xiaohongshu. UtralStrong in the bottom right corner is an app I'm currently developing; I keep it here for everyday testing.

On the computer I use Apple's built-in wallpaper, but I rarely look at the desktop. I find apps with Spotlight, put temporary files in the "Downloads" folder, then pin the folder to the Dock and find things by fanning it out. On my company's external monitor, though, I do put NetEase Cloud Music's desktop lyrics, occasionally slacking off to read the top comments.

Is there any app that you know will distract you but still can't bear to delete?

Jike. There are lots of AI product discussions there—Youmind founder Yubo, flomo co-founder Shaonan, Manus product co-founder hidecloud, and many related practitioners all share insights. But I often find myself unconsciously scrolling through it in my spare time, and what I read isn't necessarily related to AI. Even when I read a good point, putting it into practice still requires manpower, resources, and time; when I can't do it, it's easy to feel envious and develop FOMO, the fear of missing out.

For me, what matters more is solving users' problems and creating real value. Continuously using cutting-edge products also helps me maintain a certain sensitivity, so I don't have to rely on constantly scrolling through feeds to keep up.

I'm currently subscribed to flomo Max, ChatGPT Plus, Readwise, and NetEase Cloud Music VIP, and I've also bought a lifetime license for Bob Translate. The specific uses will be covered later.

How do these tools fit into your work, exercise, and creation? Has your understanding of efficiency and quality of life changed along with them?

(1) AI tools. ChatGPT Plus is my main AI tool. It can answer questions and has replaced quite a few utility apps—it's a handy digital companion. A while ago, I needed to confirm screen dimensions from hardware drawings. In the past I'd have to find software to open the drawings first, and if I couldn't understand them, I'd have to ask a hardware engineer for help interpreting them; now ChatGPT can write Python code to read the drawings, and it even reminded me to distinguish between the screen's display area and the outer bezel dimensions. For an introvert like me, it saves the step of asking others for help and eases the psychological burden.

I also use it to make small tools that reduce hassle at work. As a product manager, testing our own software and finding bugs is daily work. In the past I had to manually take screenshots, circle and annotate them, then paste them into a document to describe the problem; now, with TestFlight's screenshot feedback feature, I can circle and annotate directly and fill in descriptions, submit feedback in a centralized way, and then sync it to Feishu Bitable through a script ChatGPT helped me write. The script can also assign bugs to the corresponding colleagues based on the initials or aliases of names in the feedback, saving manual assignment and reducing the chance of errors when letting AI judge the responsible person on its own.

I also frequently use the AI built into various products. flomo's related notes and AI insights help me discover overlooked angles of thought from old notes; Reader's built-in Ghostreader can answer questions when I review reading notes, and can also connect different notes, letting content from literature and reference books echo each other. Reading experiences that were originally scattered thus become connected, which often surprises me. I have also configured a large-model translation in Bob, using it to look up the pronunciation of new words and further explain related knowledge. Simple questions can also be brought up in Bob via a shortcut, avoiding creating many temporary conversations in ChatGPT.

(2) Exercise tracking. Recently I have mainly been taking Hyrox group classes with colleagues. Unlike when I used to train for running, I don't rely as much on tools like HealthFit and Intervals.icu for feedback—after all, no matter what the data says, the training still has to be done. However, I still turn on Apple Watch recording, upholding the principle that "if you work out without wearing a watch, it means you didn't work out." Every morning I also look at my sleep data, and if I really didn't sleep well, I pause training.

When training strength alone, I mainly use the fitness app I am developing myself, letting AI generate courses while also hunting for bugs and looking for inspiration for new features.

(3) Reading and creation. I still mainly use card-based reading and writing; for the method, you can refer to "How to Take Smart Notes," while also integrating AI into it. Before starting to write, I first use flomo's AI insights to summarize old notes around the topic, to assist with drafting and reduce the psychological burden of starting from zero. If scientific viewpoints are involved, I have ChatGPT look up review papers to make a preliminary judgment on whether the viewpoint has a basis, then read the relevant papers in detail, save the content that needs close reading into Zotero, annotate while reading, and organize it into cards. Finally, combining these cards, I complete the article in Feishu or Typora.

The way I write product requirement documents is similar. After writing, I use a custom product requirement Skill to have AI check for logical errors and missed edge cases. Ideally, recurring problems and experience should also be summarized into the Skill, but I haven't done enough of this step and often forget to review and update it.

(4) Rethinking efficiency. In the past, I thought efficiency meant quickly finishing a task, ideally like a genius programmer—type a few keys and everything runs automatically. But an article written in a hurry may be full of holes, and a feature developed by AI without clearly explaining the requirements may have more than 200 bugs as soon as it is tested. Useful results, usable results, and less rework—that is the efficiency I value more now.

Pursuing efficiency should not drag you into more tasks and leave you exhausted. SSPAI's slogan is "Work efficiently, live well," and a quality life is the goal.

Therefore, I no longer rush to finish card notes, but pay attention to wording and typos and review old notes. I have also unfollowed almost all WeChat official accounts, moved my reading to WeChat Reading and Reader, and check them as needed. In the past I was always afraid of missing industry updates, but in fact I haven't missed anything important: sufficiently important news often reappears in group chats or Moments. With fewer distractions, I have also quieted down and have time to carefully experience the work and life of the present.

(5) Back to offline. This may be a bit off-topic, but one important attempt of mine this year was to solve problems again in traditional ways. At the beginning of the year, when writing my year-end summary, faced with experiences such as our parents meeting each other, getting our marriage certificate, holding two weddings, resigning and starting a new job, I didn't know where to start. After picking up pen and paper, I wrote it out smoothly instead. Then I entered it using flomo's speech-to-text, and while reading aloud I also conveniently revised the awkward parts.

When I encounter professional questions, I also ask friends directly and listen to their real experience. For example, when asking about the difference between a direct-drive and a wheel-on bike trainer, AI said thousands of words, but a friend directly told me that the wheel-on one tests balance more and cannot improve power, which let me grasp what he considered the most critical difference.

As a sports and health product manager, what changes have you observed that AI has brought to offline fitness and fitness apps? With the lower development threshold, has commercialization also become easier?

The impact on fitness apps is much greater. Ordinary people can now make a usable app within a week; with more participants, competition is fiercer, but I still don't hear of many products that can be continuously updated and bring developers revenue.

At the beginning of the year, I also spent a week using Vibe Coding to make an AI coach, but to improve the features and promote it, it might still take half a year. Materials are also a cost: the licensing fee for GIFs of 800 basic movements is about 10,000 yuan. I tried generating images with ChatGPT; for 800 movements it would be close to 1,000 yuan, and the results were not beautiful enough—many needed rework. If each GIF uses at least three images, the cost would be about 3,000 yuan. Development has become easier, but making money through fitness apps is not necessarily easier.

AI has also changed how features are implemented. Five years ago, when I was doing dietary nutrition analysis at a watch manufacturer, I had to find food databases containing nutritional information and also match images one by one—time-consuming and labor-intensive. At the time, this kind of product was mainly offered by a few companies such as Boohee in China and MyFitnessPal and Yazio abroad. Now, large models can directly provide nutritional information, reducing the work of building databases, and nutrition analysis is becoming more and more common. But the output may be unstable: for the same steamed bun, today it says 200 kcal, tomorrow it says 250 kcal. Products need to reuse users' saved settings as much as possible to keep records consistent. Developers' focus is also shifting from building databases in advance to improving the experience based on user needs.

In the future, business models may also change. Traditional fitness apps mainly sell plans and template classes—could they instead sell AI Skills? Bloggers or coaches package their experience into Skills, and after users purchase them, personalized guidance is generated, making up for the shortcomings of template classes. The same app paired with different Skills may also cover different needs such as running, strength training, and even heavy-load hiking.

How do you learn professional knowledge and improve your problem-solving ability? What role does each learning channel play?

Right now I mainly learn by doing things, improving my abilities by solving specific problems. When acquiring knowledge, I am still most accustomed to reading, though the scope of "books" is very broad and also includes Xiaohongshu—for example, when I learn swimming, I refer to content there.

Short videos tend to be fragmented, while Bilibili long videos and podcasts can provide more in-depth discussion. I often listen to "Zhixing Xiajiuguan," Meng Yan's "Nobody Knows," and "Yanzhong Huashu" hosted by Lu Yu. The first two make me think about investing in money, cognition, and personal growth, and about how investing serves life. Meng Yan has many thoughts on life, and he also showed me that leaders do not have to be hot-tempered or take pride in working overtime. Although this content does not directly teach sports knowledge, it has influenced how I work and how I treat people.

The women in "Yanzhong Huashu" showed me how to have my own garden in a flexible and self-reliant way. The sentence "Being a woman is a situation" moved me deeply. I dislike bullying, paranoia, and elite narratives in the workplace, yet I also unconsciously fall into them; these women's experiences showed me other solutions and gave me inspiration and strength.

Long videos and podcasts are still not necessarily systematic, so I need to return to books for systematic organization. Studying cognitive psychology gave me a preliminary understanding of how thinking and emotions work; "The Innovation Algorithm" made me realize that innovation also has methods; and "The Boundaries of Experience" reminded me that watching videos and reading books are both learning from others' experience, and experience has its own boundaries.

Therefore, this year I place more emphasis on firsthand experience: learning Pilates and feeling how it differs from yoga; learning swimming and experiencing how to follow the power of water. Reading, practice, asking others for advice, and my own insights complement one another; no single path can do everything. When one has not accumulated much, reading is a lower-cost way to improve oneself; once one has a certain foundation and conditions, going out to experience the world and make friends can also provide inspiration from others' experiences.