Welcome 🌟

As the title suggests, this article is a hands-on review of my first experience with the Apple Vision Pro. I spent about an hour at the Apple Eaton Centre (with the Vision Pro actually on my head for about 30 to 40 minutes), exploring many aspects of the device alongside an Apple staff member. After the whole process, I have a lot I want to share. After the actual experience, my conclusions haven’t changed compared to watching videos from other tech reviewers. As always, I habitually place the core message and my conclusions at the very beginning of the article. My current personal view is: The Vision Pro is a forward-looking spatial computing device, and its original design philosophy leaves it with no competitors in the current market. However, putting aside specific use cases and viewing it purely as a consumer electronic product, it is unacceptable at its current price. Even after all this time, it still has too many pressing issues that need to be resolved.

I recently purchased the ROG Ally (referred to as Ally below). My motivation for buying it was simple: early adoption. I wanted to see what the resulting product looks like when the PC platform’s massive game library merges with the portable “handheld” form factor. After using the ROG Ally for a week, I feel ready to organize and share my user experience.

This is an ongoing article that will be continuously updated. Its content will expand and evolve as my learning progresses. The purpose is to help myself summarize what I’ve learned about photography, color grading, and editing, while also sharing these insights with you. As someone with zero foundational knowledge but a deep passion for recording the world through a lens, I hope to approach photography and post-production with the same attitude I had when I took my first steps toward becoming a music producer. I intend to keep learning, practicing, and making infinite progress.

I have had the Alienware 18 Area-51 for a month now, and I have used it for all sorts of tasks. Standing at this point in early November 2025, this article aims to discuss two things from a technical perspective. First: How will graphics processing technology continue to evolve? Second: Compared to traditional desktop tower PCs, who exactly is the target audience for PCs in the form of heavy gaming laptops?

Following my previous article on how to build a static personal website, today I’m bringing you another technical guide: how to resolve theme conflicts when adding math formula support to a personal website. First off, Hugo’s official documentation already explains how to implement math rendering very clearly. You can find their official guide here . Since their documentation is highly detailed and covers configuration code for various file formats, I won’t repeat their steps here (though if anyone needs a Chinese breakdown, I can write one later). I will jump straight into my own troubleshooting process.

The foundational knowledges of linear algebras that needed to understand multivariable statistics techniques such as PCA, ICA and so on.

I recently bought a new Windows PC and quickly realized that the workflow is vastly different from macOS. Many of the tools I heavily relied on are simply not available on Windows. However, after installing several third-party applications, many areas I previously considered “hopeless” have seen significant improvement. I feel this list is worth sharing. First up is the input method. I chose WeChat Input . The current version has no ads, and you can enable offline mode to restrict it to local access only. The installer is only about 5 MB, so I assume it is a fairly lightweight application. In my personal experience, its phrase prediction and emoji support feel much closer to the native macOS input method. It works significantly better for my needs than the built-in Windows option.

“Absolute Mode”: Reduce language-model verbosity and keep responses concise. In my view, the prompt below helps the model answer exactly what was asked without proactively extending the topic or adding unnecessary guidance. After applying it, the model becomes more “cold” but also more efficient. If you want emotional support or casual conversation, you may prefer not to enable this mode.

A week before catching up on Evangelion, I also happened to finish reading Haruki Murakami’s novel Norwegian Wood. Why I suddenly started reading literature on a whim is a story for another time. But more than that, I am surprised I actually managed to watch Eva from beginning to end. In Norwegian Wood, Watanabe tries to understand Naoko, dedicating effort and emotion to her, yet still fails to change anything. In Eva, Shinji desperately longs to be accepted by others, but when given the chance to eliminate the A.T. Field and put an end to everything causing him pain, he refuses. Asuka also reminds me of someone from my past. What that friend did back then was as irredeemable as what I did, and I prefer not to bring it up again. But that didn’t stop me from finally understanding the reasons behind it when Eva triggered those memories five or six years later.

January 10, 2026 - LMA 0.0.3 Update If there are still people who can’t install the DMG, then I really have a bone to pick with Apple. For this update log, I asked Gemini to ghostwrite it for me; I’m done writing for now. I’ll add some actual development notes in the future. I’m exhausted—there’s just been too much going on lately. Ah… really, switching from Python to other languages takes some serious getting used to. Now that the DMG has been notarized by Apple, the system should only prompt you that it was downloaded from the internet and shouldn’t outright block you anymore. I stake my reputation on the software’s safety, and I’m hoping Gatekeeper will finally let me off the hook this time.