<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: yaqi Chen</title>
    <description>The latest articles on DEV Community by yaqi Chen (@chenchenyaqi).</description>
    <link>https://dev.to/chenchenyaqi</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3995314%2Ff40ce747-e043-4efc-8c37-7c0f52f19f80.jpg</url>
      <title>DEV Community: yaqi Chen</title>
      <link>https://dev.to/chenchenyaqi</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/chenchenyaqi"/>
    <language>en</language>
    <item>
      <title>I kept forgetting everything I learned, so I made my AI assistant actually teach me</title>
      <dc:creator>yaqi Chen</dc:creator>
      <pubDate>Sun, 21 Jun 2026 13:20:12 +0000</pubDate>
      <link>https://dev.to/chenchenyaqi/i-kept-forgetting-everything-i-learned-so-i-made-my-ai-assistant-actually-teach-me-14md</link>
      <guid>https://dev.to/chenchenyaqi/i-kept-forgetting-everything-i-learned-so-i-made-my-ai-assistant-actually-teach-me-14md</guid>
      <description>&lt;p&gt;I’ll be honest.&lt;br&gt;
You know when you spend an evening learning something new—maybe a React hook, maybe how Docker networking actually works—and it all clicks? You feel like you’ve got it. Then two weeks later someone brings it up, and it’s just… gone. Like you never touched it.&lt;/p&gt;

&lt;p&gt;That was me. Every time.&lt;br&gt;
The ridiculous part is, I have an AI coding assistant open all day, every day. It writes code, fixes bugs, answers questions. But it never taught me anything. It just handed me answers, and handed answers don’t stick.&lt;/p&gt;

&lt;p&gt;So I got annoyed and built something.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fhj8fdprlkth1nxnc451a.jpeg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fhj8fdprlkth1nxnc451a.jpeg" alt="Five Learning Skills" width="800" height="386"&gt;&lt;/a&gt;&lt;br&gt;
It’s a small CLI called Learn Anything. It takes the AI assistant you’re already using and gives it a set of teaching behaviours. Instead of “here’s the answer, bye,” it makes you understand things and remember them.&lt;/p&gt;

&lt;p&gt;It works with 30+ tools: Claude Code, Cursor, Codex, Gemini, OpenCode, Windsurf, and pretty much anything else. Run one command and your assistant suddenly has five new tricks for actually teaching you.&lt;/p&gt;

&lt;p&gt;The things it solves (because I ran into all of them)&lt;/p&gt;

&lt;p&gt;First, not knowing what you don’t know. Self-learning without a map is disorienting—random topics in random order, no sense of the whole picture. The /learn:topic command makes the AI build a knowledge map for whatever you’re studying. You can see each sub-area, what you’ve covered, what’s left, and decide what to tackle next.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjod841s7b6s8ol1xikb5.jpeg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fjod841s7b6s8ol1xikb5.jpeg" alt="Learning Knowledge Map" width="800" height="647"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Second, the AI just giving you answers. I can’t count how many times I’ve copied an AI response without really reading it. The /learn:explain command takes a different approach: it uses a recursive, Socratic style. It doesn’t dump information on you. It asks questions, lets you go deeper when you’re curious, and makes you do the thinking. Want to go three levels deep on how the event loop actually works? It’ll go there with you.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F87if3jxk10564h5ozevg.jpeg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2F87if3jxk10564h5ozevg.jpeg" alt="Deep Explanation" width="800" height="412"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Third, reading without doing. You can read about riding a bike forever and still not be able to ride one. /learn:practice drops you into TDD-style coding exercises, from warm-ups to harder challenges. You write code, the AI reviews it and gives structured feedback. That kind of repetition sticks.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fizw8r2xf2wwo1yotsow5.jpeg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fizw8r2xf2wwo1yotsow5.jpeg" alt="Practice in Time" width="800" height="450"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Fourth, the classic: forgetting everything. /learn:review uses spaced repetition, tracking where you’re weak and surfacing things right when you’re about to lose them. There’s also /learn:status, which gives you a heatmap of your mastery so you can literally see the gaps.&lt;/p&gt;

&lt;p&gt;There’s even a local dashboard. Running learn-anything serve starts a web interface where you can browse knowledge maps, go through session notes, and revisit exercises with syntax highlighting. No config, no extra dependencies.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fa080fj91zpsv2rj6pcvc.jpeg" class="article-body-image-wrapper"&gt;&lt;img src="https://media2.dev.to/dynamic/image/width=800%2Cheight=%2Cfit=scale-down%2Cgravity=auto%2Cformat=auto/https%3A%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Farticles%2Fa080fj91zpsv2rj6pcvc.jpeg" alt="Learning Dashboard" width="799" height="477"&gt;&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;If you want to try it:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npx learn-anything-cli init
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;It’ll detect your AI tools and set things up.&lt;/p&gt;

&lt;p&gt;I’m building this in the open and it improves every week. If you’re self-taught and tired of forgetting what you just learned, I’d love for you to give it a shot and tell me what sucks (and what doesn’t).&lt;/p&gt;

&lt;p&gt;GitHub: &lt;a href="https://github.com/ChenChenyaqi/learn-anything" rel="noopener noreferrer"&gt;https://github.com/ChenChenyaqi/learn-anything&lt;/a&gt;&lt;br&gt;
npm: learn-anything-cli&lt;br&gt;
A star on GitHub genuinely helps keep this going. If you have an idea or hit a bug, open an issue—I read all of them.&lt;/p&gt;

</description>
      <category>ai</category>
      <category>programming</category>
      <category>tutorial</category>
      <category>beginners</category>
    </item>
  </channel>
</rss>
