My Current Favorites – [July 2026]: My First Software, The Dream of the Red Chamber

Happiness is doing something for fun, spending hours on it, without getting paid 😀 Of course, if it paid, that would be even better, but if you enjoy an activity so much you could spend all day on it without pay, for me, that’s THE happiness.

The Dream of the Red Chamber

Lately, I’ve had almost unlimited AI use to translate whatever I want. I really liked the novel The Dream of the Red Chamber and as I told you last month, I reread the novel, and I thought maybe I should get into Redology to better understand the book. I sent tons of analyses/books/videos on the subject to the AI (Claude Code / Codex CLI) to have them translated, and thanks to those analyses, I finally understood the subtext, the wordplay, and I’m enjoying the novel 10x more.

I wrote and translated several analyses of the novel, you can read them here:

• Comparison: The Dream of the Red Chamber and In Search of Lost Time
• Determinism or free will?
• Episode 1: Cao Xueqin, the man and the book (I) – Zhou Ruchang
• Episode 2: Cao Xueqin, the man and the book (II) – Zhou Ruchang
• Episode 4: How was The Dream of the Red Chamber written? (I)
• Episode 5: How was The Dream of the Red Chamber written? (II)
• etc.

One of the novel’s themes is that fate sends us plenty of opportunities, plenty of warnings, to help us redirect our path. But most people find it incredibly hard to leave the road they’re already on. Giving up what you have for an uncertain future, or stopping and saying you’ve already received enough, is extremely difficult. In a world where everyone always wants ‘more’, ‘better’, choosing ‘less’, ‘little’ goes against the grain.

The novel also warns us against the danger of putting personal interest ahead of the common good. We see developed countries heading straight for the wall (looking at you, France!), we can clearly see where that path leads, and yet nothing changes, because the warnings go unheeded.

Codex CLI

For several weeks now, thanks to OpenAI’s generosity in handing out seemingly endless tokens to users, I’ve been taking on increasingly interesting AI projects.

I probably would never have started increasingly complex projects without nearly unlimited access to AI. It really shows one thing: when you have abundance, you take more risks. On a large scale, the risk seems almost nonexistent, so you dare more. It’s a bit like billionaires who can try many things without truly endangering their lives, whereas an employee who leaves a permanent contract to start their own business takes a real risk. The more leeway you have, the more you can afford to fail (I talked about it here)

I’ve also been encouraged and inspired by Shania Khan, she worked in marketing like me, and released two paid iPhone apps written by AI. I figured there was plenty for non-developers to do.

---

I’ll just tell you about three cool projects I did, two of which are for reMarkable. My tablet/e-reader attracts a lot of geeks, it has SSH access, so you can create many tools to interact with it.

Project 1: I had the AI build a web interface to extract highlights from reMarkable and send them to ANKI so I can review them from time to time. You have ready-made prompts here if you want and the GitHub repo here.

This interface is available on your phone via the local network.

Project 2: I created a tool to change the screensaver on my e-reader. We could have done it via command line or Filezilla, but a web interface is nicer, especially since you can zoom/crop the photo as you like.

Project 3: a simple text to speech to listen to my books out loud, a free Audible alternative.

• Why? Most books aren’t on Audible and sometimes I want to listen to articles (websites) out loud too.
• How? This simple tool uses Edge & Microsoft voices (free) and I set it up to read in three languages (the ones I speak). The Vietnamese voices are actually the most natural, despite a Southern accent. The tool works on both computer and smartphone, locally with a shared library, you can add books as well as websites. You can pause and resume a book at any time.
• Who is it for? I think this tool is ideal if you often have your hands full or if you have parents with poor eyesight.

On smartphone, the behavior is very similar to Audible 😀

Then, since my parents are interested in this program and setting up the web interface is a bit tedious, I thought “I can do better than that”, I vibe-coded a Windows application that they can install on their computer. As you can see, the tool has become a combined Audible & Kindle. It’s available in three languages (FR-EN-VI)

This is my very first software ! I’m so thrilled! I found it really more difficult than all the other projects combined, because you have to work on UX, UI, there are tons of small features to add; plus security issues to handle (epubs and pdfs can be malware nests). Anyway, I can’t believe I created my first Windows software!

---

For me, this kind of tool (Claude Code, Codex CLI) is Internet 2.0. We are at a time when 99% of humanity still doesn’t know how to use it fully, so if you start now, you’ll have a head start and access to opportunities. The Internet enriched a lot of people, AI will do the same.

More info: I tested Claude Code and you should too and my tips for beginners.

I think we all have a lot of opportunities. Sure, we can’t buy apartments at ridiculously low prices like our parents or grandparents did, but in exchange, we have other cool opportunities like the Internet, social networks, AI. We also have opportunities to apologize, to make amends for our wrongs; when I listen to the stories of elderly people, I realize that we (the universe) have always left us plenty of moments to (1) improve our karma (2) push us toward our “personal legend” – but we don’t necessarily see them, and we miss them.

How to use AI to be happier

For me, everyone could try to:

1. create an AI copy of yourself to work less and enjoy life
2. then, create an improved AI version of yourself (that does things you can’t/don’t know how to/don’t want to do) to be even more efficient and work even less
3. and if possible, reach a level where you create a product that generates passive income and become FIRE

Finding projects to do is a skill. Inventing projects for yourself is also activating your creativity.

• For point (1), I take the example of someone who does the same reporting every week. They go fetch info from a certain place, copy-paste it into PowerPoint, do the same analyses. Might as well automate all that to save time and avoid the risk of copy-pasting the wrong data.
• For point (2), I take the example of a teacher. To work better, they can start from a standard assignment, make 25 versions, so that each student receives a version at their level. For the top student, the demonstration of a theorem for example will just have one sentence “prove that …”, and the bottom student will have 36,000 suggestions to reach the demonstration. This way, it gives the right exercise to the right student, so the strongest doesn’t get bored, and the weakest doesn’t give up. To do this, you can feed the assignment to Codex CLI or Claude Code, list all the students’ names and specify each student’s level (very good / good /… poor), and the AI will adapt the standard assignment into 25 versions tailored to each. In two minutes, it will produce 25 PDF assignments with each student’s name to print, without taking too much of the teacher’s time.
• For point (3), I take the example of an influencer who sells “shopping guides”: it’s a small list of winter coats with good composition and good value for money, I think she sells it for €20. This list, when done by hand, is really long and boring. But she can clearly automate it with AI: a crawler that looks at about thirty brands she usually checks, reads the compositions, compares prices, and selects the best clothes. The influencer just reviews the list to check style, color… and gives feedback so the AI improves next time. Twelve times a year, she can put out her “shopping guide” spending very little time on it (or even no time at all).

Knowing how to use these AI coding tools will at least make you competitive. “The barrier to coding is lower today than it has ever been. People who code, whether they are CEOs, marketers, recruiters, and not just software engineers, will accomplish far more than those who don’t code.”, said Andrew Ng, co-founder of Coursera, founder of DeepLearning.AI, former head of Google Brain. The most likely effect of AI is not a total elimination of jobs, but a transformation of tasks and skills, with a productivity gain for those who know how to use these tools.

If you need an AI trainer for your company, contact JB ! JB might be able to help you. The idea is to train each expert in your company to use AI so they can create their own tool – to go faster and make bullshit jobs less boring. JB was a developer, he’ll know how to explain it better than me.

The right timing is NOW

In the near future (a matter of a few years), those who know how to use AI will gain a significant edge over others: they’ll go faster, further thanks to AI while freeing up free time. But that edge will be temporary. The rest of the world will gradually catch up with early adopters, the general pace will accelerate, and that gap that allowed you to breathe will disappear. We’ll all become more efficient, without being able to capitalize on that edge sustainably. That’s exactly why you need to get into it now: the window during which this edge constitutes a real advantage is limited, the longer you wait, the more you let it close without having taken advantage of it.

For businesses, AI will make industries even more efficient, capable of offering personalized services at lower cost, whether for individuals (custom clothing, shoes…) or for companies (developing an in-house CRM tailored exactly to their needs without paying for Salesforce, HubSpot, or Zendesk licenses). This democratization will directly weaken the model of large consulting firms and solution providers, who will no longer be able to justify their high fees. They will now have to prove real added value and rare expertise. This will lead to a redistribution of wealth : established players who lived off their dominant position will have to reinvent themselves, while smaller, more agile structures will be able to compete thanks to AI in markets that were previously closed to them.

For younger generations, especially future graduates, the window is gradually closing, but it remains open. AI isn’t adopted everywhere yet, leaving them time to build a solid knowledge base the traditional way. Because before asking AI to code, you need to know what an algorithm is, how to code by hand—you don’t necessarily have time to learn all that once you’re on the job. So the basics must be acquired before automating. This is also a key period for developing emotional intelligence, intuition, critical judgment, the ability to navigate uncertainty, and the ability to make decisions—things AI doesn’t have.