The Creation of the Ask AI Plugin Was Accidental
A little story about this plugin.
In January 2025, I tried reading through a Python tutorial book twice, but still struggled to even begin to grasp the basics. I even filled a dedicated Python Learning folder with file after file of code snippets: I tried again to build my Python code foundation by breaking down larger, more macro-level modules into smaller, more granular learning and operational tasks (based on each code snippet file). Of course, I could pessimistically and clearly foresee what would happen a month later: even with the increasing number of code files in the folder, I knew that after a month, with my current textual and cognitive tools for different modules, I still wouldn't be able to string together enough complete code chains to build a small product.
Later, throughout the process, I continuously searched for new tools, especially those that could provide more transparency and foundational knowledge during the learning process. Until I rediscovered the AI thinking mode that could crack the "code chain," I realized that when you can not only see "someone else's" solution, but also see the thought process behind their answer to a question at each step, even if that "someone else" is an AI.
With the support of "more transparent and explanatory AI products," I understood and mastered more operational common sense, and wrote a very small product called Notifier: Utilizing system push notifications, it uses a JSON table containing words and translations as the title and notification content of a notification, sending them randomly at set times to reinforce memory. This was a very interesting productization attempt. Given time, if combined with Anki's memory concept and expanded from a JSON data set of only 10 words and translations, it could become a standalone product.
Then, AI IDE products exploded in popularity. I was surprised to find that although the way AI generates code, the process, and even the results are somewhat wild, if you describe your requirements in sufficient detail (extremely fine-grained), your familiarity with the code will increase significantly. Later, after several months of iteration, AI IDE products, through various automated tools, have made a seemingly random AI increasingly proficient, and the transparency of the entire process for specific tasks has improved.
Finally, the Ask Grok plugin was launched and iterated to a stable Ask AI Plugin version.
Isn't Reading the Same?
Books are like a codebase with a slower iteration speed, but they have stricter review standards, professional editors, and market-driven verification and screening. Readers attempt to unravel mysteries, shape their own thinking tools, and construct their own unique database of experiences using ample data. In this shaping process, how to accelerate understanding, how to map and transform experiences from elsewhere—perhaps from the past—into the present environment, abstracting and refining new wisdom from one's own unique contextual information, isn't this also the path of human "emergent wisdom"?
AI cannot provide all the correct answers, but it can offer the possibility of accelerating the unraveling of mysteries. You are welcome to use the Ask AI Plugin to explore any mysteries in the knowledge base.
The Ask AI Plugin is still being updated; this is my first open-source project, but it won't be the last.