Mekal Z

Mekal Z

A programmer running out of the wall.

Weekly Share #21, 2023

First, manually add a header image using Midjourney.


Cool things I did#

A bit of a messy rhythm recently#

The rhythm of life and work in the past couple of months has been a bit chaotic. I used to keep a daily review for a long time, recording my daily journal in Logseq, but I haven't been recording much in the past couple of months.
I'm used to recording what I did each day, what books, videos, or articles I read, interesting things I discovered, people I met, and plans for tomorrow, etc. Basically, anything can be recorded. This process is actually a way to organize and free up my brain. After recording many things in writing, my brain doesn't have to worry about remembering them anymore. When I need them, I can simply filter through the double-linked tags or do a full-text search.
In fact, when I write down what I was thinking at the time or what I'm currently reading, it's a process of forcing my brain to think and summarize further. For example, when I read an article and record it in Logseq, I subconsciously think about several questions:

  1. How should I categorize this article?
  2. How is this article related to what I'm currently doing?
  3. What similar articles have I read before?
  4. What is the main content of this article?
  5. What did I gain from this article?
  6. If I were to share it with someone else, how should I introduce it?

Through these thoughts, the article leaves a deeper impression in my mind, and I can quickly review related topics or knowledge points at this time.

I have set up two main categories for my Daily Journal template: Plan/Unplanned. Whenever I record anything, I need to think about whether it is part of my plan, so I always consider whether I am focusing on important things. Through continuous feedback, I try to stay focused as much as possible.

For example, here is a record from a day in March when I was still conceptualizing the "Linggua Vocabulary" product 😂:


In the past, when I did weekly reviews and wrote this Weekly Share, I only needed to filter through some tags to easily obtain a large amount of material from the past week. So doing reviews and writing blog posts felt easy, and I was more willing to do them.
But recently, due to laziness and a disrupted rhythm, I haven't been consistent with this journal, which makes it difficult for me to recall what I did this week when writing the weekly review/share. This state can be described as "muddled". When it comes to reviewing and writing blog posts on weekends, I find it even more difficult, so I keep procrastinating. That's why the Weekly Share has been interrupted for several weeks recently.

"Linggua Vocabulary"#

Fixed some minor issues and optimized several performance problems. At the same time, I created a WeChat group, hoping that friends who want to communicate with me or other English learners can join by scanning the code.

Introduction and download link for "Linggua Vocabulary":


Texts I wrote#

It seems that I just posted some answers on Zhihu.

Videos I Watch#

I rewatched the pronunciation tutorial by "English Rabbit". By the way, "English Rabbit" is the best English teacher I have ever seen, no doubt about it. Highly recommended!

I watched a lot of English learning videos by Tiffani. I also highly recommend her. She speaks standard American English, very easy to understand, and very enthusiastic.

French Singer Alizée#

This is a French singer I followed many years ago. She was really popular at that time, but unfortunately, her glory has faded away.
Here are a couple of videos from before. The video quality is not great, but she sings well and dances well. You'll definitely want to watch them multiple times!

Cool stuff I discovered#

Comparing the effectiveness of different large language models#

If you need to determine which large language models are suitable for your use case and want to compare the effects of different models, you can try this website:
It has three modes:

  1. Single Model
    In this mode, you can choose a model to experiment with. It supports many models:


  1. ⚔️ Chatbot Arena ⚔️ (Battle)
    In this mode, you can input a prompt and two models that you don't know in advance will provide results.
  2. ⚔️ Chatbot Arena ⚔️ (Side by Side)
    In this mode, you can choose two different models to provide feedback on the same prompt, and then you can compare the results.

Introducing a few more large language model products#

Ownership of this post data is guaranteed by blockchain and smart contracts to the creator alone.