<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
    <channel>
        <title>Learning on Rebel Zhang&#39;s Blog</title>
        <link>https://rebel1725.codeberg.page/blog/en/tags/learning/</link>
        <description>Recent content in Learning on Rebel Zhang&#39;s Blog</description>
        <generator>Hugo -- gohugo.io</generator>
        <language>en-GB</language>
        <managingEditor>rebel1725@tilde.club (Rebel Zhang)</managingEditor>
        <webMaster>rebel1725@tilde.club (Rebel Zhang)</webMaster>
        <lastBuildDate>Wed, 11 Nov 2026 21:09:19 +0800</lastBuildDate><atom:link href="https://rebel1725.codeberg.page/blog/en/tags/learning/index.xml" rel="self" type="application/rss+xml" /><item>
            <title>A Very Useful Learning Approach</title>
            <link>https://rebel1725.codeberg.page/blog/en/post/a-very-useful-learning-approach/</link>
            <pubDate>Wed, 11 Nov 2026 21:09:19 +0800</pubDate><author>rebel1725@tilde.club (Rebel Zhang)</author>
            <guid>https://rebel1725.codeberg.page/blog/en/post/a-very-useful-learning-approach/</guid>
            <description>&lt;p&gt;&lt;strong&gt;TL;DR: Try explaining your thoughts to someone else, whether a friend or an LLM. Ask them to evaluate your reasoning and discuss your ideas with you. This is likely to reveal misunderstandings and help you learn something new.&lt;/strong&gt;&lt;/p&gt;&#xA;&lt;p&gt;A few things have happened recently:&lt;/p&gt;&#xA;&lt;ol&gt;&#xA;&lt;li&gt;The physical-world version of me had a conversation with one of his classmates about CPU scheduling. When asked about HRRN (High Response Ratio Next), the question reminded him to check whether SJF (Shortest Job First) is pre-emptive. He initially thought it was, but soon checked and found that it is not.&lt;/li&gt;&#xA;&lt;li&gt;Over the past few days, I have been studying advanced C programming, starting with ownership and lifetimes. After developing an initial understanding, I wrote a blog post and sent it to an LLM for grammar and spelling corrections. However, it spotted a serious &lt;em&gt;technical&lt;/em&gt; issue in my text, and further investigation showed that my understanding of lifetimes was incorrect.&lt;/li&gt;&#xA;&lt;li&gt;While writing this blog post, I also refined my understanding of SJF and SRTF: scheduling is based on the length of the next CPU burst, rather than the total amount of CPU time remaining across all future bursts. For example, suppose job A has a 5 ms CPU burst, followed by an I/O operation and then a 7 ms CPU burst, while job B has a single 8 ms CPU burst. The scheduler should choose A rather than B.&lt;/li&gt;&#xA;&lt;/ol&gt;&#xA;&lt;p&gt;These experiences have taught me something important about learning.&lt;/p&gt;&#xA;&lt;p&gt;When I learn a new concept, I sometimes form assumptions or intuitions that I have not yet verified but nevertheless treat as correct, even though they may be wrong. For example, when I first learnt about SJF, I understood it to mean that the job with the shortest remaining execution time gets the CPU, and that if a job with a shorter remaining time joins the ready queue, it pre-empts the currently running process. The discussion mentioned above helped me realise that this is not how SJF works. After discussing it with an LLM, I also learnt that &amp;lsquo;shortest remaining execution time&amp;rsquo; is not the right criterion for describing non-pre-emptive SJF: SJF selects the process with the shortest next CPU burst, whereas SRTF selects the process with the shortest remaining time in its current CPU burst.&lt;/p&gt;&#xA;&lt;p&gt;To consolidate your understanding of a concept, I suggest explaining it in your own words to someone else. You could start by explaining it to an LLM. Doing so gives you a way to test whether your understanding is sound; if your explanation contains mistakes, the LLM may point them out. However, do not assume that every correction from an LLM is right: check important claims against reliable sources or other explanations. If you are looking for an LLM service that places an emphasis on privacy and anonymity, &lt;a class=&#34;link&#34; href=&#34;https://duck.ai/&#34;  target=&#34;_blank&#34; rel=&#34;noopener&#34;&#xA;    &gt;Duck.ai&lt;/a&gt; is one option.&lt;/p&gt;&#xA;&lt;p&gt;If you have friends who are learning the same concept, discussing your ideas with them is another good option. Each of you may have different misconceptions, and you can help one another correct them. Furthermore, a problem your friend is struggling with may draw your attention to a gap in your own understanding that you had not noticed (as in the first example above), and vice versa.&lt;/p&gt;&#xA;&lt;p&gt;Exchanging ideas with your friends can also broaden your thinking and knowledge, and may lead you to interesting topics you have not yet explored.&lt;/p&gt;&#xA;&lt;p&gt;Learning together can be much more effective and efficient than learning alone. This is one reason why &lt;a class=&#34;link&#34; href=&#34;https://gnu.org/philosophy/free-sw.html&#34;  target=&#34;_blank&#34; rel=&#34;noopener&#34;&#xA;    &gt;free software&lt;/a&gt; and &lt;a class=&#34;link&#34; href=&#34;https://creativecommons.org/&#34;  target=&#34;_blank&#34; rel=&#34;noopener&#34;&#xA;    &gt;free culture&lt;/a&gt; are important, and one reason why the free software community matters: when free software users, developers and enthusiasts exchange ideas, everyone learns and benefits.&lt;/p&gt;&#xA;</description>
        </item></channel>
</rss>
