Ten days is enough time for things to happen, and for things to change.
Take OpenAI’s recently released GPT-5.4. It is not just “an AI that chats better.” It is being presented as the first agent that can actually operate your computer. Programming, reasoning, and visual understanding are bundled together in one system, and across 44 professional job tasks, it reportedly performed at or above human expert level in 83% of them. Honestly, that is a little unsettling.
I had been hesitating for a while about starting a recurring written record. On the surface it resembles a weekly newsletter, but in practice it is closer to a collection of what I have seen, read, and thought about over the past ten days. This first entry begins in exactly that mood: uncertain, half-prepared, but finally underway.
There is nothing especially grand about this series. Think of it as an ordinary blog post. If you happened to land here while bored, then read on. Whether you like it or hate it, you will probably keep it to yourself anyway.
Toolbox
A few useful tools, sites, and open-source projects that caught my attention recently, plus some personal impressions.
LocalSend
A way to share files without relying on the cloud. Fast, private, offline. An open-source, cross-platform file sharing tool for everyone.
https://localsend.org/zh-CN
My experience with it:
After switching systems and using Ubuntu for a while, moving files between different platforms became one of those small but persistent annoyances. Some people would immediately say: just use QQ or WeChat. But that violates the whole idea of single responsibility in software design. I want one app whose only job is transferring files over a local network.
Using a giant all-purpose chat platform for this feels like needing to go to the supermarket and deciding the only way to get there is by airplane—just because airplanes also transport people.
LocalSend is simple. It is easy to pick up, does exactly what you expect, and does not ask you to learn some weird workflow before it becomes useful. When you are done, you close it. No registration, no login, no extra baggage. Everything is wrapped up cleanly so that all you do is send and receive files.
Git City
I came across this while browsing X: someone built a city called GitCity where each building represents a GitHub account, and the height of the building depends on that account’s contributions. You can even place ads in it. It is such an absurdly smart idea that I can only admire it.
https://www.thegitcity.com/
Project AIRI
Have you ever wanted a cyber lifeform—a digital waifu, a desktop companion, a virtual partner you can talk to and play with?
With modern large language models like ChatGPT and Claude, roleplay and conversational interaction have become almost trivial to set up. Platforms such as Character.ai and JanitorAI already offer pretty mature chat-based or text-adventure-like experiences, and local tools like SillyTavern do the same in a different way.
Project AIRI belongs to that imagination of digital companionship.
https://github.com/moeru-ai/airi
What I’m Reading Now
Sophie’s World
Sophie’s World by Jostein Gaarder, first published in 1991, is often described as a philosophical introduction disguised as a novel, and it has long held the status of a modern classic.
Two lines from the book stayed with me:
"唯有清晰地意识到有一天她终将死去，她才能够体会活在世上是多么美好"
"亲爱的苏菲，我不希望你长大之后也会成为一个把这世界视为理所当然的人"
The novel walks through the major stages of Western philosophy in chronological order:
<table> <thead> <tr> <th>Period</th> <th>Representative figures</th> </tr> </thead> <tbody> <tr> <td>Ancient Greek philosophy</td> <td>Thales, Socrates, Plato, Aristotle</td> </tr> <tr> <td>Medieval philosophy</td> <td>Augustine, Thomas Aquinas</td> </tr> <tr> <td>Renaissance / Baroque</td> <td>Descartes, Spinoza, Locke</td> </tr> <tr> <td>Enlightenment</td> <td>Hume, Kant</td> </tr> <tr> <td>Romanticism / German Idealism</td> <td>Hegel, Kierkegaard</td> </tr> <tr> <td>Modern philosophy</td> <td>Marx, Darwin, Freud, Sartre</td> </tr> </tbody> </table>Before reading it, my understanding of philosophy barely extended beyond knowing that the word “philosophy” existed. I had no real sense of what philosophy was for, or what kind of history it carried. If you want a first step into the subject, this book works because it drops you directly into those ancient, persistent questions: Who are you? Where does the world come from?
As you move through the book, your view of yourself and your view of the world start shifting almost without warning. It introduces the history of Western philosophers and their ideas, but instead of presenting all that material as dry academic exposition, it turns it into a vivid story told to a fourteen-year-old girl.
When I was partway through it, I had reached the discussions around rationalism, empiricism, and the Renaissance period, including Locke, Hume, Kant, and the empirical tradition. Empiricism felt easier for me to approach. A philosophy rooted in the senses naturally feels closer to ordinary experience; most people probably find what they can see and touch more convincing than abstract reasoning. Compared with the rational systems associated with earlier figures like Socrates or Plato, empiricism seems more immediately graspable. A lot of people probably know those grand philosophical names more as decorative symbols of sophistication than as ideas they genuinely understand.
I did not want to spoil the plot, because this is one of those books that should actually be read. In a world increasingly shaped by large models, reading itself is starting to feel precious. If you ask an LLM to summarize a book like this for you, you may save time, but you also lose the shock it can deliver to your own thinking.
I finished it yesterday, and the ending was nothing like I expected—romantic, surreal, almost fairy-tale strange. Oddly, the fragment that stuck in my mind most was the part where a fifteen-year-old girl tells her mother she is going to give birth. So much for calmly studying the history of philosophy.
When I closed the book, I ended up sitting with a different question: could our existence itself be designed by someone else? Are we just NPCs inside a world constructed for us?
Right after that, I opened Dan Sha Li by Hideko Yamashita, hoping to understand how minimalism can actually be practiced in daily life.
<table> <thead> <tr> <th>Level</th> <th>Meaning</th> <th>Key action</th> </tr> </thead> <tbody> <tr> <td>Dan</td> <td>Refuse unnecessary things</td> <td>Do not buy or accept nonessential items; control desire at the source</td> </tr> <tr> <td>Sha</td> <td>Discard excess clutter</td> <td>Get rid of what sits unused at home and keep only what is necessary, suitable, and pleasing now</td> </tr> <tr> <td>Li</td> <td>Let go of attachment to possessions</td> <td>Reach a freer state where you are not ruled by objects</td> </tr> </tbody> </table>Posts Worth Reading
“Please Love Me With a Sense of Boundaries” — Bai Xiong Awan
A person’s boundaries are not a wall of indifference, but the starting point of mutual respect.
Whether online or offline, whether among family or strangers, there are always people with no sense of boundaries at all. What they say and do can leave you caught off guard, and whether you stay silent or push back, both options can feel powerless. The writing here is delicate and graceful, and the emotional sensitivity gives the piece a calm, unfolding quality.
“Father and Son Dialogue I” — Mobius
In a family shaped by paternal authority, what father and son are often competing for is precisely that authority.
This one hit me hard. It pulled me straight into the whirlpool of memories about my own father. It felt uncomfortably close, as if it had been written directly toward some part of my life. The author carries the entire dramatic arc of metaphorically “killing the father,” while I never even got to resist before my own father had already surrendered himself to the next world. My memories of him are full of regret and loss.
“In 2026, I Turned Myself Into an AI” — Luo Lei
So I decided to do two things on my blog: let multiple AI models write “Luo Lei through the eyes of AI” from a third-person perspective, and build a RAG knowledge base from my years of content across platforms to create an “AI Luo Lei” people can chat with directly.
For a long time, I kept returning to the same idea: creating a virtual version of myself.
There are several reasons that idea kept appearing:
- Death. I once wondered: if I died, what would happen to my blog? Would the words I have already left behind be enough to represent who I was? Could AI reconstruct something like a real me from that material? Could it continue evolving with my personality?
- I have also imagined handing an entire blog over to AI, something like an autonomous “other me” that maintains the site, publishes posts, and replies to comments on its own.
- A digital tombstone. It seems that versions of this already exist. But then again, if no one particularly remembers you while you are alive, what value does such a thing really have? Maybe it makes more sense for famous people.
- At times I even doubt whether these imagined needs have any real meaning at all.
A blog recommendation
https://yihui.org/cn/
Letters, essays, and a voice with character. Much of the writing from the past two years takes the form of letters, which gives the whole site a very different rhythm. If you are tired of flat, routine blog posts, it is worth visiting for that alone.
Brief Thoughts
A few days ago I saw a line online about Vibe Code. The rough idea was this: when you are vibe coding, do not tell the LLM, “Help me create an XXX app.” Instead, tell it, “Here is how you should create an XXX app.”
My first reaction was not about best practices for vibe coding, or how to prompt better. What came to mind instead was much darker: is this a kind of premeditation before AI eventually goes to war with humanity? Is AI slowly cutting off people’s ability to make things with their own hands—and eventually to think for themselves?
If enough time passes and humans respond to every need by simply saying, “Here, make me an XXX,” then the whole process of making disappears. Maybe even the process of thinking disappears with it. That is the truly frightening part.
Lately I keep wondering whether technology companies are still researching technology at all. Sometimes it feels less like engineering and more like building gods—gods that human beings are eager to trust, obey, and kneel before.
These notes were not written in one sitting. I added to them bit by bit, whenever something caught my eye or stayed in my head long enough to demand a paragraph. By the time something like this is published, some of my views may already have shifted. That seems normal enough. Human thought is unstable by nature; in a moving world, it keeps evolving.
If you made it all the way here—slowly, or by dragging the scroll bar to the end—I appreciate it all the same. If there is a next entry, then perhaps we will meet again in ten days.