<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Muhib on Medium]]></title>
        <description><![CDATA[Stories by Muhib on Medium]]></description>
        <link>https://medium.com/@skillsatwork.skills?source=rss-443824bd1245------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*j50YunT6SQWDC7IFaMyAuA.jpeg</url>
            <title>Stories by Muhib on Medium</title>
            <link>https://medium.com/@skillsatwork.skills?source=rss-443824bd1245------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 18:51:38 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@skillsatwork.skills/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Take Breaks like Smart People]]></title>
            <link>https://medium.com/illumination/take-breaks-like-smart-people-95da5df5deca?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/95da5df5deca</guid>
            <category><![CDATA[medicine]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[science]]></category>
            <category><![CDATA[self-improvement]]></category>
            <category><![CDATA[life-lessons]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 08:23:46 GMT</pubDate>
            <atom:updated>2026-10-29T08:23:46.302Z</atom:updated>
            <content:encoded><![CDATA[<h4>Not all breaks are equal. Try to rank yours.</h4><figure><img alt="Coffee Break" src="https://cdn-images-1.medium.com/max/1024/0*2t5BUO1PqUx-Lipt" /><figcaption>Photo by <a href="https://unsplash.com/@mindspacestudio?utm_source=medium&amp;utm_medium=referral">Mindspace Studio</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>Breaks.</strong> Aren’t they something we eagerly look forward to? They feel like a dearly awaited island on a long sea voyage. But does our brain enjoy the break equally? Does it get rejuvenated, or further tired?</p><p>A study by <strong>Jaroslaw Grobelny, Marta Glinka &amp; Teresa Chirkowska-Smolak</strong> answered these questions.</p><h3>Microbreaks</h3><p><strong>Micro breaks</strong> are brief, intentional pauses lasting less than 10 minutes. They don’t disrupt workflow, but they extend the longevity of our productivity.</p><h3>1. Nature-related micro breaks</h3><ul><li><strong>Activity:</strong> Greet your eyes with natural landscapes (forests, mountains, coastlines) through images or short videos, paired with ambient nature sounds.</li><li><strong>Result:</strong> <em>Highest overall performance</em>. Nature breaks returned both vigor and fatigue levels completely to pre-work baselines.</li><li><strong>Why it works:</strong> Engaging with nature uses indirect attention, so it adds no mental demand. It restores cognitive focus, calms the nervous system, and elevates positive affect.</li></ul><h3>2. Blank or quiet rest micro breaks</h3><ul><li><strong>Activity:</strong> A brief pause in a shrine of quietness, without any device, task, or external media.</li><li><strong>Result:</strong> <em>Outperformed digital breaks</em> in fatigue reduction, fatigue relief, and vigor restoration. Vigor restoration was comparatively greater than with social media breaks.</li><li><strong>Why it works:</strong> It halts cognitive stimulation altogether, maximizes relaxation, and gives you autonomy over your time.</li></ul><h3>3. Hedonic social media breaks</h3><ul><li><strong>Activity:</strong> Watching algorithm-served short-form content, with light interaction through likes and dislikes.</li><li><strong>Result:</strong> <em>Partial replenishment</em>. It slowed fatigue accumulation and boosted vigor, but failed to bring energy levels back to baseline.</li><li><strong>Why it works:</strong> Social media excels at psychological detachment. Immersive, algorithm-driven feeds shift focus away from stress. By and large, it is a great tool for <em>escapism</em>.</li><li><strong>Where it fails:</strong> On two fronts, relaxation and fatigue recovery. Triggering dopamine release and nervous system arousal hinders relaxation. It also consumes cognitive energy, since dynamic digital content demands direct attention and emotional processing.</li></ul><h3>4. No break (continuous work)</h3><ul><li><strong>Activity:</strong> Working through tasks without any bouts of relaxation or replenishment.</li><li><strong>Result:</strong> <em>Severe resource drain.</em> Continuous work means continuous energy depletion, so fatigue accumulates faster.</li></ul><h3>Takeaways</h3><ul><li>For stopping work thoughts, social media is an outstanding tool. Still, digital overstimulation can be taxic.</li><li>Brief social media use isn’t inherently destructive to productivity, but as a primary rest mechanism, it’s a poor choice.</li><li><strong>Ranking:</strong> Nature micro breaks are best overall. Quiet rest is best for calm and control. Continuous work is worst.</li></ul><p>Source — <a href="https://pubmed.ncbi.nlm.nih.gov/39284901/">The impact of hedonic social media use during microbreaks on employee resources recovery</a></p><p>You might also like — <a href="https://medium.com/illumination/parkinsons-law-why-work-expands-to-fill-the-time-you-give-it-d0ad0ddba563">Parkinson’s Law: Why Work Expands to Fill the Time You Give It</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=95da5df5deca" width="1" height="1" alt=""><hr><p><a href="https://medium.com/illumination/take-breaks-like-smart-people-95da5df5deca">Take Breaks like Smart People</a> was originally published in <a href="https://medium.com/illumination">ILLUMINATION</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Parkinson’s Law: Why Work Expands to Fill the Time You Give It]]></title>
            <link>https://medium.com/illumination/parkinsons-law-why-work-expands-to-fill-the-time-you-give-it-d0ad0ddba563?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/d0ad0ddba563</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[self-improvement]]></category>
            <category><![CDATA[time-management]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Sun, 27 Sep 2026 09:29:10 GMT</pubDate>
            <atom:updated>2026-09-27T09:29:10.792Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Eventually everything hits the bottom, and all you have to do is wait until someone comes along, and turns it back again. ⌛️" src="https://cdn-images-1.medium.com/max/1024/0*3sv-2Vxwg1F1UTKt" /><figcaption>Photo by <a href="https://unsplash.com/@aronvisuals?utm_source=medium&amp;utm_medium=referral">Aron Visuals</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>An elderly woman decides to write a postcard. She has an entire day to herself.</p><p>She spends an hour finding the postcard. Another hour hunting for her spectacles. Another composing the note.</p><p>Let’s do the math. A postcard holds six to eight lines, each with eight to ten words — roughly <strong>48 to 80 words</strong> total. The average handwriting speed sits around <strong>13 words per minute</strong>. At that pace, the note should take <strong>3.7 to 6.2 minutes</strong> to write. To give her the benefit of age, lets take it as 10 minutes.</p><p>But, she took the entire day.</p><p>Why? Northcote Parkinson has the answer.</p><p><strong>The Law, Defined</strong></p><p>In 1955, in <em>The Economist</em>, C. Northcote Parkinson wrote a satirical essay with a simple premise: <strong>work expands to fill the time available for its completion</strong>.</p><p>The size of the task barely matters. What determines how long something takes is the time you allocate to it. Less time forces you to move quickly. More time lets the same task stretch, stall, and idle for as long as the clock allows.</p><p><strong>From One Person to an Organization</strong></p><p>The same law scales from an individual to an entire institution. Staff numbers grow regardless of workload — a pattern Parkinson traced through Admiralty statistics. The British Navy was shrinking at the time, yet the bureaucracy managing it kept expanding. Headcount and workload, it turns out, have little to do with each other.</p><p><strong>The Law of Triviality</strong></p><p>Parkinson also noticed something else. A committee reviewing plans for a nuclear power plant spent most of its time debating the materials for a bike shed.</p><p>Time spent on an agenda item, he observed, is often inversely proportional to its actual importance — not because the bike shed mattered more, but because it was the easiest thing in the room to argue about. (The same committee that debates a font choice for twenty minutes will approve a six-figure budget in two.)</p><figure><img alt="Sharpened #2 pencils in a cup" src="https://cdn-images-1.medium.com/max/1024/0*C20am2MzGnDoPrrR" /><figcaption>Photo by <a href="https://unsplash.com/@kellysikkema?utm_source=medium&amp;utm_medium=referral">Kelly Sikkema</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><strong>Takeaways</strong></p><ol><li><strong>It’s a behavioral tendency, not a strict law: </strong>People tend to focus on the time available, not the time actually required. When time isn’t gated, effort spreads like water — expanding until it meets a barrier.</li><li><strong>A tighter deadline sharpens focus:</strong> A self-imposed time limit pushes you toward a sharper level of efficiency than you’d otherwise find on your own.</li><li><strong>We gravitate toward the easy, not the important:</strong> We clutch onto the easiest not the important and devote our precious time and attention to something that is going to play little or no role in the progress of our work.</li><li><strong>Name the task before you name the time:</strong> Estimate the real effort first — as we did with the postcard — then set the deadline. Most people do this backwards: they size the clock first and let the task expand to match it.</li><li><strong>Smaller groups make faster, better decisions:</strong> It is a version of the old adage: <strong>too many cooks spoil the broth</strong>. Parkinson found that the most effective governing bodies stayed small.</li></ol><p><strong>Conclusion</strong></p><p>You already have the postcard, the clock, and the time.</p><p>You have to decide your writing speed — the deadline you set for yourself, not the one the day hands you.</p><p><strong>For the Full Story</strong></p><p>Everything above draws from Parkinson’s original 1958 book, <a href="https://link.amazon/B07jkRUoi">Parkinson’s Law: The Pursuit of Progress</a><em>. </em>If the postcard story stuck with you, the book is worth the read.</p><p>If you want to read more about productivity frameworks and practical ways to think better in an AI-driven world, <a href="https://iamthemarginalist.substack.com/">join</a> my free newsletter.</p><p><em>Disclosure: This article contains affiliate links. If you purchase through them, I may earn a small commission at no extra cost to you.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d0ad0ddba563" width="1" height="1" alt=""><hr><p><a href="https://medium.com/illumination/parkinsons-law-why-work-expands-to-fill-the-time-you-give-it-d0ad0ddba563">Parkinson’s Law: Why Work Expands to Fill the Time You Give It</a> was originally published in <a href="https://medium.com/illumination">ILLUMINATION</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Get More by Working Less]]></title>
            <link>https://medium.com/illumination/get-more-by-working-less-ea1b4a3b3a38?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/ea1b4a3b3a38</guid>
            <category><![CDATA[books]]></category>
            <category><![CDATA[life]]></category>
            <category><![CDATA[life-lessons]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[time-management]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Sat, 15 Aug 2026 13:14:52 GMT</pubDate>
            <atom:updated>2026-08-15T13:14:52.309Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Planning Your Work" src="https://cdn-images-1.medium.com/max/1024/0*o1TeZjFXfyLjiB3p" /><figcaption>Photo by <a href="https://unsplash.com/@mattragland?utm_source=medium&amp;utm_medium=referral">Matt Ragland</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Vilfredo Pareto was an Italian economist-sociologist and Professor of Political Economy at Lausanne, Switzerland. While he was researching land and wealth distribution in Italy in the late 1800s, he calculated that roughly 80% of land belonged to about 20% of the population. He also noticed the same lopsided pattern in his garden, where 80% of the peas came from a fractional 20% of the pods.</p><p>This made him suspect that this unequal distribution was like a universal law. At last, he published this finding in his book <em>Cours d’Économie Politique</em>.</p><p>Little did he know that he would be talked about a century later.</p><h3>80–20 Rule</h3><p>By the way, Pareto never called it the 80–20 rule. It was Dr. Joseph Juran who attributed this pattern to Pareto and named it Pareto’s Principle. He argued that this wasn’t just an economic deduction but a universal principle applicable across diverse facets of life. He even went on to coin the famous phrase — “the vital few and the trivial many.”</p><h3>Why It Works</h3><p>It is an undeniable fact that most things in life are unevenly distributed, and this unevenness extends to our efforts too. That’s why 20% of our efforts or inputs tend to produce nearly 80% of outcomes.</p><p>This isn’t a strict law, rather it is a pattern — the numbers won’t always be 80 and 20. It can be 70–30 or 90–10, but what matters is that a small number of causes will always drive a large share of effects. That’s why people who manage to produce outsized results concentrate on a small set of high-leverage activities and treat the rest as busywork.</p><figure><img alt="Charting Goals" src="https://cdn-images-1.medium.com/max/1024/0*u9S7kLPmIfVp3OdN" /><figcaption>Photo by <a href="https://unsplash.com/@isaacmsmith?utm_source=medium&amp;utm_medium=referral">Isaac Smith</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Turning It Into a Productivity Habit</h3><p>The writing on the wall is very straightforward — focus on the vital few and stop spreading attention evenly across all your tasks. To put it into action:</p><ol><li><strong>Audit before you optimize</strong> — Take out last month’s data and rank your actions by their impact. It could be revenue generated for a product, problems solved, or relationships strengthened.</li><li><strong>Find your 20%</strong> — Based on the ranking, pick out the high-impact actions. They will make up your 20%. It almost always turns out to be a very short list.</li><li><strong>Protect that 20% ruthlessly</strong> — That 20% is like gold dust to you, as it forms the backbone of your work. Shield it valiantly from interruptions and distractions.</li><li><strong>Deliberately shrink the 80%</strong> — The remaining 80% isn’t useless, but it isn’t of utmost importance either; it falls somewhere in between. You can automate it, batch it, or cut it, so it doesn’t steal away the precious time allotted to your dear 20%.</li></ol><h3>Honest Caveat</h3><p>The Pareto Principle is not a foolproof scientific law. It has its downsides — it has been criticized for being more a product of statistical convenience than a genuine universal law. If used carelessly, it can push people toward short-term wins at the expense of long-term stability. It must be used more as a lens for prioritizing than a formula to plug values into.</p><h3>Takeaway</h3><p>The real advantage of the 80/20 rule isn’t the exact ratio — it’s the implicit question it forces you to ask: <em>what is actually driving me, or my work, forward?</em></p><p>This question must find a secure place in your strategy before you take the driver’s seat and drive into the world of possibilities.</p><p>📘 <strong>Want the original source?</strong> Richard Koch’s <a href="https://link.amazon/B03QDFE03">The 80/20 Principle</a> talks about this concept in detail. You can read and find more ways to apply in your work and life.</p><p>If you want to read more about productivity frameworks and practical ways to think better in an AI-driven world, join my free newsletter.</p><p><strong>[</strong><a href="https://iamthemarginalist.substack.com/"><strong>Subscribe here</strong></a><strong>]</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ea1b4a3b3a38" width="1" height="1" alt=""><hr><p><a href="https://medium.com/illumination/get-more-by-working-less-ea1b4a3b3a38">Get More by Working Less</a> was originally published in <a href="https://medium.com/illumination">ILLUMINATION</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[GLM-5.2 — A Breath of Fresh Air in the AI Atmosphere]]></title>
            <link>https://ai.plainenglish.io/glm-5-2-a-breath-of-fresh-air-in-the-ai-atmosphere-321a8484a13e?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/321a8484a13e</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[chatgpt]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Sat, 11 Jul 2026 07:20:59 GMT</pubDate>
            <atom:updated>2026-07-11T07:20:59.031Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="3D render of AI and GPU processors" src="https://cdn-images-1.medium.com/max/1024/0*gXAchw3G8h1bN-OX" /><figcaption>Photo by <a href="https://unsplash.com/@omilaev?utm_source=medium&amp;utm_medium=referral">Igor Omilaev</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p><em>A Chinese university lab just built one of the best AI models of 2026 — matching ChatGPT and Claude, and offering it at a fraction of the price.</em></p><p>You must have heard the story of David and Goliath. The AI theatre is witnessing its remake, reiterating the beauty of stories — that they remain relevant in any timeline.</p><p>Tsinghua University, the Chinese equivalent of MIT, started a research project in 2021. Five years later, in 2026, they have created David — GLM-5.2 — that is looking Goliath in the eye. That Goliath being Claude and ChatGPT. Not only that, in coding, it is craning its head over the giant. Independent benchmarks are already calling GLM-5.2 the best open source AI model of 2026.</p><p><strong>What is GLM-5.2?</strong></p><p>GLM, General Language Model, is a new AI launched by Z.ai in June 2026. It is particularly built for coding and agentic work — that is, rather than a one-time answer, it can plan, execute, and reiterate on complex multi-step tasks on its own. The newcomer gate-crashed the closed-door race of AI by outsmarting its competitors.</p><h4>What makes it special?</h4><ul><li>It is extremely capable</li><li>Its free tier matches the best paid models</li><li>It is open source AI with an MIT license</li><li>It is cheap to access via API</li><li>It has a one million token context window — a differentiator</li></ul><p><strong>Why is GLM-5.2’s One Million Token Context Window Making News?</strong></p><p>Let’s understand the concept of tokens. One million tokens is roughly 750,000 words. It is equivalent to ten novels. Every AI has a limited context window, that is, a defined word limit within which it retains memory, and beyond which it falters. For most AI models like ChatGPT and Claude, it is around 100,000 tokens. This makes GLM-5.2 a Gulliver in front of Lilliput natives.</p><p>It has practical implications. For example, a software developer can paste their entire code and GLM-5.2 will be able to reason across all of it at once, without forgetting anything. This is huge. This is setting standards.</p><p><strong>MIT License — What Open Source AI in 2026 Looks Like</strong></p><p>Most AI models are like a vending machine — you pay, they provide, but you never get to see the machine. Well, time is up for that machine. GLM-5.2 is open source AI. You can download the code, use it, modify it, and make products from it without paying a single penny. This is what open source AI in 2026 looks like — free, modifiable, and commercially usable by anyone.</p><p>The MIT license — the most generous of all — does away with royalties, revenue thresholds, and commercial restrictions.</p><p><strong>GLM-5.2 Pricing vs ChatGPT and Claude</strong></p><p>AI models are charged by the token — a token being roughly three-quarters of a word. Here is what the market looks like right now:</p><ul><li>Claude Opus 4.8 (Anthropic): $25 per million output tokens</li><li>GPT-5.5 (OpenAI): $30 per million output tokens</li><li>GLM-5.2 (Z.ai): $4.40 per million output tokens</li></ul><p>Yes. GLM-5.2 is disrupting the market. It is five to seven times cheaper than established AI models.</p><p><strong>What Does This Mean?</strong></p><p>I foresee two changes in the AI market, courtesy of GLM-5.2. First, prices will drop. With a frontier-quality model delivering at such minimal cost, its competitors will certainly follow the lead. Second, AI models from regions other than America will dare to dream. We might see a few others making it to the international pedestal.</p><p>🤖 <strong>Fun Fact</strong></p><p>Ask GLM-5.2 who it is, and it will sometimes say — Claude. If you try to correct it, then also it stays firm as the Rock of Gibraltar in its belief. The widespread explanation says that, Z.ai trained their model on massive internet data, a significant chunk of which was Claude’s responses. This Claude-tinged training left an impression on its identity. It is just a quirk, not a flaw as its reasoning and code are entirely its own.</p><p>Try out GLM-5.2 with hands-on experience, as the battle of David and Goliath goes on in the background.</p><p>Click here to know how I use ChatGPT to harvest its maximum potential.</p><p><a href="https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3">https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=321a8484a13e" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/glm-5-2-a-breath-of-fresh-air-in-the-ai-atmosphere-321a8484a13e">GLM-5.2 — A Breath of Fresh Air in the AI Atmosphere</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[ChatGPT Shortcuts Aren’t Secret Codes — Here’s Where They Actually Came From]]></title>
            <link>https://medium.com/illumination/chatgpt-shortcuts-arent-secret-codes-here-s-where-they-actually-came-from-da12067890e4?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/da12067890e4</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[writing-prompts]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[chatgpt]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Fri, 10 Jul 2026 03:32:50 GMT</pubDate>
            <atom:updated>2026-07-10T03:32:50.825Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="chatgpt-shortcuts-explained.png" src="https://cdn-images-1.medium.com/max/1024/0*yt_xdVfaLg_4-cky" /><figcaption>Photo by <a href="https://unsplash.com/@boliviainteligente?utm_source=medium&amp;utm_medium=referral">BoliviaInteligente</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Cheat codes for AI, just like a GTA game? No. Not actually.</p><p>Our web is a witness to dozens of blogs and articles labelling conventional and not-so-conventional words as “cheat codes.” But they are essentially one-word or short instructions to AI models, dressed up as <strong>ChatGPT shortcuts</strong>.</p><p>ChatGPT doesn’t decode a phrase like “TL;DR” the way a computer decodes a keyboard shortcut. It understands the phrase. It derives that understanding from us — so in a way, it understands just like any other human would. An online visitor knows TL;DR means “summarize this.” It means the same thing to the model.</p><p>While some of these <strong>ChatGPT shortcuts</strong> have consistent results, others yield non-uniform responses. Read ahead to find out which is which.</p><h3>The Most Common ChatGPT Shortcuts</h3><h4><strong>ELI5 / ELI10</strong></h4><p><em>Nomenclature:</em> In September 2011, a Reddit user named bossgalaga started the subreddit r/explainlikeimfive. A month later, ELI5 entered Urban Dictionary. Two months after that, it appeared on Twitter. Soon users started prefacing questions with the condensed form — ELI5.</p><p>ELI10 is an unofficial variant — same idea, a little less baby talk.</p><p><em>Example:</em> “ELI10: How does photosynthesis occur?”</p><p>It’ll explain it to you in the simplest terms, treating you like a ten-year-old. The complex process gets melted into plain language, easy to comprehend.</p><p>I just wonder how a five-year-old would actually use this.</p><h4><strong>TL;DR</strong></h4><p><em>Nomenclature:</em> Dates back to 2002 — a Usenet post on rec.games.video.nintendo. A year later, it entered Urban Dictionary. Originally it was used to dismiss someone’s long post. Soon, people started including it in their own writing, to summarize for readers who wouldn’t read the whole thing.</p><p><em>Example:</em> Paste a long article, add “TL;DR” at the end.</p><p>A short summary of a few sentences greets you.</p><h4><strong>JARGONIZE</strong></h4><p><em>Nomenclature:</em> Dates two centuries back, to 1803. Used to mean “make into jargon” or “express in jargon.” Not a slang , but a dictionary-standard word repurposed for prompting. <a href="https://www.merriam-webster.com/dictionary/jargonize">Source — Merraim Webster.</a></p><p><em>Example:</em> “JARGONIZE: The computer is slow because too many tabs are open.”</p><p>Returns something like: “The system is experiencing performance degradation due to excessive concurrent process allocation.”</p><h4><strong>LISTIFY</strong></h4><p><em>Nomenclature:</em> A modern coinage, devoid of slang history and dictionary entry — unlike SIMPLIFY. Built by adding the suffix “-ify” to the word “list,” specifically for AI prompting.</p><p><em>Example:</em> Paste a messy meeting transcript, add “LISTIFY.”</p><p>A clean, bullet-pointed list of action items gets returned, rummaged out of the mess.</p><h4><strong>SIMPLIFY</strong></h4><p><em>Nomenclature:</em> A standard English word with no AI-era backstory. Means what it has always meant.</p><p><em>Example:</em> “SIMPLIFY: The aforementioned clause stipulates non-negotiable terms of engagement” becomes “This rule can’t be changed.”</p><h4><strong>CRITIQUE</strong></h4><p><em>Nomenclature:</em> Long before it became a prompt code, it was used for structured evaluation.</p><p><em>Example:</em> Paste your own cover letter, add “CRITIQUE.” It puts down specific weak points, not just praise.</p><p>Not praising is quite unbecoming of AI.</p><h4><strong>ANALOGIZE</strong></h4><p><em>Nomenclature:</em> Another ordinary dictionary word — the verb form of “analogy.”</p><p><em>Example:</em> “ANALOGIZE: How does a blockchain work?”</p><p>You’ll likely get something like a shared ledger everyone can see, but no one can secretly edit.</p><h4><strong>HUMANIZE</strong></h4><p><em>Nomenclature:</em> Usage dates back to the 1600s — used to describe making something more human, natural, humane. Now it’s widely used to make AI sound less like a machine.</p><p><em>Example:</em> Paste AI-generated copy, add “HUMANIZE.”</p><p>It comes back sounding like a person wrote it, not a manual.</p><h4>ACT AS [role]</h4><p><em>Nomenclature:</em> More than a word — a prompting technique where you assign a character to the model, which improves the response in multiple ways. It emerged alongside chatbots themselves, not borrowed from older language.</p><p><em>Example:</em> “ACT AS a senior editor. Critique this paragraph.”</p><p>It returns a sharper, more specific feedback than a plain critique request.</p><p><a href="https://ai.plainenglish.io/the-3-word-change-that-makes-chatgpt-actually-useful-65cf416ffbbe">Read more about this prompting technique.</a></p><h3>ChatGPT Shortcuts That Don’t Actually Work</h3><h4><strong>MUDA</strong></h4><p><em>Nomenclature:</em> A core concept in the Toyota Production System for “futility” or “wastefulness.” This Japanese word has no history as an AI term. Though imported into prompting, it requires defining in the same prompt to work reliably.</p><p><em>Example:</em> “MUDA — identify wasted steps in this workflow: [paste process]” works. “MUDA” on its own, undefined, gives inconsistent results.</p><h4><strong>QOE</strong></h4><p><em>Nomenclature:</em> It is a standard business-consulting shorthand, borrowed loosely from finance. In prompting it has no fixed meaning — it’s repurposed as “identify non-value-adding tasks.”</p><p><em>Example:</em> Works only with context: “QOE — flag steps in this process that don’t add value: [paste process].”</p><h4>VISUALIZE#</h4><p><em>Nomenclature:</em> Not a real term. It is undefined — just a made-up placeholder that was never actually defined by whoever coined it.</p><p><em>Example:</em> “VISUALIZE#: sales data” gives you a guess. “Present this as a table with columns for month, region, revenue” gives you exactly what you want.</p><h3>The Real Lesson</h3><p>Every code with a genuine linguistic history works, full stop — the AI learned it the same way it learned everything else, from us. Others, with no regard for history, need context spelled out in the same prompt to be understood.</p><p>I’d prefer to build my own self-explanatory commands than memorize someone else’s cheat sheet.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=da12067890e4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/illumination/chatgpt-shortcuts-arent-secret-codes-here-s-where-they-actually-came-from-da12067890e4">ChatGPT Shortcuts Aren’t Secret Codes — Here’s Where They Actually Came From</a> was originally published in <a href="https://medium.com/illumination">ILLUMINATION</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[GLM-5.2 — A Breath of Fresh Air in the AI Atmosphere]]></title>
            <link>https://medium.com/@skillsatwork.skills/glm-5-2-a-breath-of-fresh-air-in-the-ai-atmosphere-c4b049d64ff2?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/c4b049d64ff2</guid>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Fri, 03 Jul 2026 03:53:11 GMT</pubDate>
            <atom:updated>2026-07-03T03:54:09.698Z</atom:updated>
            <content:encoded><![CDATA[<p><em>A Chinese university lab just built one of the best AI models of 2026 — matching ChatGPT and Claude, and offering it at a fraction of the price.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*Gq7p__Vk4KKiRXoK" /><figcaption>Photo by <a href="https://unsplash.com/@steve_j?utm_source=medium&amp;utm_medium=referral">Steve A Johnson</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>You must have heard the story of David and Goliath. The AI theatre is witnessing its remake, reiterating the beauty of stories — that they remain relevant in any timeline.</p><p>Tsinghua University, the Chinese equivalent of MIT, started a research project in 2021. Five years later, in 2026, they have created David — GLM-5.2 — that is looking Goliath in the eye. That Goliath being Claude and ChatGPT. Not only that, in coding, it is craning its head over the giant. Independent benchmarks are already calling GLM-5.2 the best open source AI model of 2026.</p><p><strong>What is GLM-5.2?</strong></p><p>GLM, General Language Model, is a new AI launched by Z.ai in June 2026. It is particularly built for coding and agentic work — that is, rather than a one-time answer, it can plan, execute, and reiterate on complex multi-step tasks on its own. The newcomer gate-crashed the closed-door race of AI by outsmarting its competitors.</p><h4>What makes it special?</h4><ul><li>It is extremely capable</li><li>Its free tier matches the best paid models</li><li>It is open source AI with an MIT license</li><li>It is cheap to access via API</li><li>It has a one million token context window — a differentiator</li></ul><p><strong>Why is GLM-5.2’s One Million Token Context Window Making News?</strong></p><p>Let’s understand the concept of tokens. One million tokens is roughly 750,000 words. It is equivalent to ten novels. Every AI has a limited context window, that is, a defined word limit within which it retains memory, and beyond which it falters. For most AI models like ChatGPT and Claude, it is around 100,000 tokens. This makes GLM-5.2 a Gulliver in front of Lilliput natives.</p><p>It has practical implications. For example, a software developer can paste their entire code and GLM-5.2 will be able to reason across all of it at once, without forgetting anything. This is huge. This is setting standards.</p><p><strong>MIT License — What Open Source AI in 2026 Looks Like</strong></p><p>Most AI models are like a vending machine — you pay, they provide, but you never get to see the machine. Well, time is up for that machine. GLM-5.2 is open source AI. You can download the code, use it, modify it, and make products from it without paying a single penny. This is what open source AI in 2026 looks like — free, modifiable, and commercially usable by anyone.</p><p>The MIT license — the most generous of all — does away with royalties, revenue thresholds, and commercial restrictions.</p><p><strong>GLM-5.2 Pricing vs ChatGPT and Claude</strong></p><p>AI models are charged by the token — a token being roughly three-quarters of a word. Here is what the market looks like right now:</p><ul><li>Claude Opus 4.8 (Anthropic): $25 per million output tokens</li><li>GPT-5.5 (OpenAI): $30 per million output tokens</li><li>GLM-5.2 (Z.ai): $4.40 per million output tokens</li></ul><p>Yes. GLM-5.2 is disrupting the market. It is five to seven times cheaper than established AI models.</p><p><strong>What Does This Mean?</strong></p><p>I foresee two changes in the AI market, courtesy of GLM-5.2. First, prices will drop. With a frontier-quality model delivering at such minimal cost, its competitors will certainly follow the lead. Second, AI models from regions other than America will dare to dream. We might see a few others making it to the international pedestal.</p><p>🤖 <strong>Fun Fact</strong></p><p>Ask GLM-5.2 who it is, and it will sometimes say — Claude. If you try to correct it, then also it stays firm as the Rock of Gibraltar in its belief. The widespread explanation says that, Z.ai trained their model on massive internet data, a significant chunk of which was Claude’s responses. This Claude-tinged training left an impression on its identity. It is just a quirk, not a flaw as its reasoning and code are entirely its own.</p><p>Try out GLM-5.2 with hands-on experience, as the battle of David and Goliath goes on in the background.</p><p>Click here to know how I use ChatGPT to harvest its maximum potential.</p><p><a href="https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3">https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c4b049d64ff2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[I Changed 3 Words in My ChatGPT Prompts. The Difference Was Embarrassing]]></title>
            <link>https://ai.plainenglish.io/the-3-word-change-that-makes-chatgpt-actually-useful-65cf416ffbbe?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/65cf416ffbbe</guid>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[writing-prompts]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Wed, 29 Apr 2026 10:44:43 GMT</pubDate>
            <atom:updated>2026-04-30T00:22:36.527Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*D307da_MiLiOzb9u" /><figcaption>Photo by <a href="https://unsplash.com/@berctk?utm_source=medium&amp;utm_medium=referral">Berke Citak</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><blockquote>I almost gave up on ChatGPT. Then I found this.</blockquote><p>Let me be upfront about something.</p><p>Three months ago, I thought ChatGPT was genuinely overrated. Not in a contrarian way — I just wasn&#39;t getting anything useful out of it. I&#39;d ask something, get a wall of text back, read it, and think &quot;okay but this doesn&#39;t actually help me.&quot;</p><p>My friends were acting like this thing had changed their lives. On the other hand, I was getting responses that sounded like they were written by someone who had read a lot about human beings but never actually met one.</p><p>So I nearly gave up on it entirely.</p><p>I didn&#39;t, obviously. And the thing that changed my mind was so small it&#39;s almost annoying to admit.</p><h3>It’s Three Words</h3><p>“Act as a...&quot;</p><p>That&#39;s it. As simple as that.</p><p>I know how that sounds. Bear with me for two more minutes because the difference it makes is not subtle — it&#39;s the kind of thing where you&#39;ll try it once and immediately feel stupid for not knowing it sooner.</p><p>Here&#39;s the problem with how most people use ChatGPT. You open it, you type your question, you hit enter.</p><p>What you&#39;re essentially doing is asking a question to... nobody in particular. ChatGPT has no idea if you&#39;re a student, a professional, a complete beginner, or someone with twenty years of experience. So it hedges. It writes something that technically answers the question for literally anyone who might ask it.</p><p>Which is why it always feels slightly off. It&#39;s not written for you. It&#39;s written for the average of every possible person who could have asked that question.</p><p>The moment you say &quot;act as a...&quot; you give it a specific lens to look through. And everything that comes out the other side is different.</p><h3>Here’s What Changed For Me — Actual Examples</h3><p>I don&#39;t want to just explain this in theory because, honestly, that&#39;s boring. Let me show you what actually happened when I started using this.</p><h4>First time I tried it properly:</h4><p>I was studying and needed to understand something complicated. The Krebs cycle — it&#39;s a biochemistry concept that looks like someone spilled spaghetti on a whiteboard.</p><p>The old way, I would have asked:</p><blockquote>&quot;Explain the Krebs Cycle.&quot;</blockquote><p>What I got: a paragraph that read like the first result on Google. Accurate, I suppose. Completely failed to land in my brain in any meaningful way.</p><p>New prompt:</p><blockquote>&quot;Act as a medical professor who’s really good at explaining things to students who are new to biochemistry. Use an analogy. Make it memorable.&quot;</blockquote><p>What came back was actually good. Like, genuinely good. It used an analogy about an energy factory, walked through the stages in a way that made logical sense, and ended with a tip for remembering the sequence. I read it once, and it stuck.</p><p>Same question. Completely different output.</p><h4>Second example — a work situation:</h4><p>I had to email someone about a late payment. Awkward situation. I didn&#39;t want to be aggressive, but I also couldn&#39;t just pretend it wasn&#39;t happening.</p><p>Asked ChatGPT normally: &quot;Help me write a professional email about a missed payment.&quot;</p><p>Got back: a template with [CLIENT NAME] and [AMOUNT] in brackets. The kind of thing that would make anyone feel like they received a form letter.</p><p>Useless.</p><p>Tried again with a better prompt:</p><blockquote>&quot;Act as someone really good at professional communication. I need to email a client who missed a payment. I want to be direct but not cold — I still want to keep the relationship intact. Keep it short, under 100 words.&quot;</blockquote><p>What it wrote — I actually sent it. Nearly word for word. It was measured, clear and human. Did the job without creating more awkwardness.</p><h4>Third one — content ideas:</h4><blockquote>&quot;Act as a content strategist who works specifically with writers on Medium covering AI and productivity. My target readers are students and young professionals. Give me article ideas that aren’t the same five things everyone else is writing about.&quot;</blockquote><p>First time I&#39;d gotten content suggestions that didn&#39;t make me immediately think &quot;I&#39;ve seen this exact article seventeen times this week.&quot;</p><h3>Why Does This Actually Work Though</h3><p>It is fairly simple.</p><p>It&#39;s because vague inputs produce vague outputs — and most of us ask vague questions without realising it. When you assign a role, you&#39;re not just changing the style of the answer. You&#39;re changing the assumptions the model makes about what you already know, what level of depth you need, and what kind of answer would actually be useful to you.</p><p>It&#39;s the difference between asking a random stranger for directions versus asking someone who grew up in that neighbourhood. The information they draw from is the same city. But one of them knows which roads have construction and which shortcut actually saves time.</p><h3>What more can be done</h3><p>You can push this further, too. The role is one layer. Adding context about yourself is the second layer. Adding a constraint is the third.</p><p>Just &quot;act as a fitness coach&quot; is decent.</p><blockquote>&quot;Act as a fitness coach helping someone who&#39;s 25 years old, works a desk job, has 20 minutes a day maximum and a lower back problem they&#39;re managing&quot; — that&#39;s a completely different level of specificity. And the response reflects that.</blockquote><p>It takes maybe ten extra seconds to write. That&#39;s the entire tradeoff.</p><h3>The Honest Limitations</h3><p>I&#39;d feel weird writing this without saying where it falls flat.</p><p>If your underlying question is too vague, the role framing helps, but doesn&#39;t completely rescue it. You still need to know roughly what you want.</p><p>If you lack clarity in your question, the same cloudiness would be reflected in the answer.</p><p>It does nothing for real-time information. ChatGPT&#39;s knowledge has a cutoff. No role you assign changes that.</p><p>And occasionally it&#39;ll agree to the role and then quietly drift away from it after a few sentences. When that happens, just add &quot;maintain this role throughout your entire response&quot; at the end of your prompt. Fixes it almost every time.</p><h3>The Hack</h3><p>If you are too lazy to write the prompt yourself, ask ChatGPT to write it for you. If you provide a decent context about what you are asking, then you can expect an above-average response.</p><p>It works almost all the time.</p><figure><img alt="scrabble, scrabble pieces, lettering, letters, wood, scrabble tiles, white background, words, quote, letters, type, typography, design, layout, focus, bokeh, blur, photography, images, image, try, go for it, just do it, don&#39;t give up, keep going, one foot in front of the other, if not now when, procrastination, determination, motivation, " src="https://cdn-images-1.medium.com/max/1024/0*pkFUbJZcY3Nfzw_b" /><figcaption>Photo by <a href="https://unsplash.com/@brett_jordan?utm_source=medium&amp;utm_medium=referral">Brett Jordan</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>Just Try It Once</h3><p>Genuinely — go find something you asked ChatGPT before and weren&#39;t happy with. Paste the same question, but put &quot;act as a [relevant expert]&quot; at the front and add one sentence about your situation.</p><p>See what comes back.</p><p>I&#39;ve shown this to people who&#39;ve been using ChatGPT for over a year. The reaction is always some version of mild frustration that nobody mentioned it earlier.</p><p>Which is exactly how I felt. Hence, this article.</p><p>I also trained my ChatGPT to respond way better than others. <a href="https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3">Read here</a>.</p><h3>A message from our Founder</h3><p>Hey, <a href="https://linkedin.com/in/sunilsandhu">Sunil</a> here. I wanted to take a moment to thank you for reading until the end and for being a part of this community. Did you know that our team run these publications as a volunteer effort to over 3.5m monthly readers? We don’t receive any funding, we do this to support the community.</p><p>If you want to show some love, please take a moment to follow me on <a href="https://linkedin.com/in/sunilsandhu">LinkedIn</a>, <a href="https://tiktok.com/@messyfounder">TikTok</a>, <a href="https://instagram.com/sunilsandhu">Instagram</a>. You can also subscribe to our <a href="https://newsletter.plainenglish.io/">weekly newsletter</a>. And before you go, don’t forget to clap and follow the writer️!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=65cf416ffbbe" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/the-3-word-change-that-makes-chatgpt-actually-useful-65cf416ffbbe">I Changed 3 Words in My ChatGPT Prompts. The Difference Was Embarrassing</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[My ChatGPT Responds Better Than Yours, Here is the 3-Step Guide]]></title>
            <link>https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/05e25c57c5f3</guid>
            <category><![CDATA[education]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[prompt-engineering]]></category>
            <category><![CDATA[medicine]]></category>
            <category><![CDATA[productivity]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Tue, 21 Apr 2026 23:03:28 GMT</pubDate>
            <atom:updated>2026-04-21T23:03:28.840Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*ARuxcNOQO3-MJc3Y" /><figcaption>Photo by <a href="https://unsplash.com/@solomin?utm_source=medium&amp;utm_medium=referral">Dima Solomin</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>We can train our AI to generate the best response. A response tailor made for our queries. It will be in perfect alignment with the context.<br>As a student, I use AI as a study companion. I subscribed to many of those as horses for courses. For quick answers, I always resort to ChatGPT. My ChatGPT responds differently and more efficiently than my friends. Why? Because I trained it for immaculate responses.</p><h3>The Project</h3><p>I assign a project for one subject. For example, I would create separate projects for Forensic Medicine, Preventive &amp; Social Medicine, Ophthalmology, etc. Within the project, you have options to add one or multiple sources and instructions. You can write instructions directly here or in the chat section. Both ways it works fine. <br>Personally, I put the instructions as a first message in the chat because this allows me to instruct my AI better by letting it ask questions.</p><h3>Let AI ask questions</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/799/1*1_ce4m4ECpoTfUy1ALnlHQ.jpeg" /><figcaption>Image by Author</figcaption></figure><p>We are accustomed to blasting questions at AI with the anticipation of getting the desired response all the time. Sometimes, it succeeds and other times it fails.<br>AI is your companion. This implicitly suggest that your relationship with AI is ot one-way. So questions can be asked by AI as well.<br>When I give a set of instructions, I always ask AI whether it wants to ask something. This way, my instructions and directions get better and more precise. This is reflected in the responses generated following the instructions.</p><h3>My Instruction Template</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*WU14SYURaBm4cfaMV-clag.jpeg" /><figcaption>Image by Author</figcaption></figure><p>I assign a character to my AI. I list the sources from which it has to extract information. Then, I tend to outline all the possible scenarios with it, such as classroom cross-question, viva question, exam question, etc. Then, I assign a slash command to each of these scenarios.</p><h3>My Slash Commands</h3><p>Having an arsenal of slash commands is a boon. Not only does it save time, but it also makes the entire process seamless. <br><strong>/c</strong> — classroom cross-questions. Answers should be short and precise without any jargon.<br><strong>/v</strong> — viva questions. Answers should be short and precise without any jargon. <br><strong>/e</strong> — exam questions. Answers should be complete with bullet points. It must align with my medical university guidelines. <br><strong>/5</strong> — Answer in 5 lines.</p><h3>My Workflow</h3><p>Now, I have my companion ready. It is at my beck and call.<br>If my strict professor asks a tough question in class and I don’t know the answer, then I will immediately hit the “/c&quot; command.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*9Za5maC4Z6HeVCQTX-R-mQ.jpeg" /><figcaption>Image by Author</figcaption></figure><p>If I am standing outside the examination hall waiting for my viva, my batchmate exiting from the hall reveals a question asked to him. Without any panic and hassle, I will hit the “/v&quot; command, which will save me from losing a mark or two.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*TIqkE2pQYq-nzolG3VTd7g.jpeg" /><figcaption>Image by Author</figcaption></figure><p>If I am preparing for my exams, then I can get answers for all the important questions in a perfect format by using the “/e&quot; command.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/799/1*XqODIemHZgGF_vVBBvzopA.jpeg" /><figcaption>Image by Author</figcaption></figure><p>AI has a problem of generating lengthy responses. Even for one-word answer, it vomits an enormous amount of text. At times, it becomes frustrating. I fixed it with the “/5&quot; command. It restricts the answer to just 5 lines.</p><h3>Conclusion</h3><p>We all have access to AI, but few people know how to use it for their work. The peculiar features of Project and Slash commands make CharGPT an outstanding study tool. With a little bit of training, we can transform a generic AI into high performance tool. The aim is not to get an answer but a correct answer, that too when needed. This aim can be realised by following this guide.</p><blockquote>Which slash command are you planning to use?</blockquote><h3>A message from our Founder</h3><p>Hey, <a href="https://linkedin.com/in/sunilsandhu">Sunil</a> here. I wanted to take a moment to thank you for reading until the end and for being a part of this community. Did you know that our team run these publications as a volunteer effort to over 3.5m monthly readers? We don’t receive any funding, we do this to support the community.</p><p>If you want to show some love, please take a moment to follow me on <a href="https://linkedin.com/in/sunilsandhu">LinkedIn</a>, <a href="https://tiktok.com/@messyfounder">TikTok</a>, <a href="https://instagram.com/sunilsandhu">Instagram</a>. You can also subscribe to our <a href="https://newsletter.plainenglish.io/">weekly newsletter</a>. And before you go, don’t forget to clap and follow the writer️!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=05e25c57c5f3" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/my-chatgpt-responds-better-than-yours-here-is-the-3-step-guide-05e25c57c5f3">My ChatGPT Responds Better Than Yours, Here is the 3-Step Guide</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[7 Ways to Use Excel as a Productivity System in 2026]]></title>
            <link>https://medium.com/@skillsatwork.skills/7-ways-to-use-excel-as-a-productivity-system-in-2026-091c90913d3f?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/091c90913d3f</guid>
            <category><![CDATA[study]]></category>
            <category><![CDATA[time-management]]></category>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[business]]></category>
            <category><![CDATA[excel]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Mon, 13 Apr 2026 15:47:16 GMT</pubDate>
            <atom:updated>2026-07-21T01:13:51.736Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="An iced coffee helping with the hot weather" src="https://cdn-images-1.medium.com/max/1024/0*xyWXZf0q77KJMj1Z" /><figcaption>Photo by <a href="https://unsplash.com/@_ferh97?utm_source=medium&amp;utm_medium=referral">Fernando Hernandez</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Excel is for Accountants and Professionals.</p><p>This is the belief of the general public, but it is not true. Aside from professionals, we can also use it for tracking and auditing our personal lives. It is highly customizable to fit our needs. It is free and available offline — a rarity in this world where everything needs an internet connection to run.</p><p>There are tons of productivity apps available, but they have their caveats. They are expensive, collect data, live in separate places, and often make our lives more complicated than simple.</p><p>Unlike these apps, Excel allows you to keep everything under one roof. Your budget, tasks, habits, and goals are safely housed in one app with easy control. This tremendously useful app can be used for the following purposes:</p><h3>1. Budget Tracking — Know Where Every Rupee Goes</h3><p>Most people have no idea where their money goes. This is an expensive liability. For such people, Excel is the best option.</p><p>It has built-in templates that can be put to use. A little knowledge about spreadsheets can help you create your own tracker that is specifically designed for you.</p><p>Besides, it will be free, without recurring subscription fees and syncing discrepancies.</p><h3>2. Task and Project Management — Without the Subscription Fee</h3><p>This free tool can manage your tasks and projects just like paid apps. You just need to play with a handful of columns like task name, deadline, priority, and status. Once the base is laid, add conditional formatting to flag deadlines in red.</p><p>You can also build Gantt charts for multiple projects using standard bar charts.</p><h3>3. Habit Tracking — Build Consistency With a Grid</h3><p>Healthy habits like reading books and journaling can transform a person completely. Since you cannot improve what you cannot measure, a tool is essential. Excel ticks all the boxes for such a tool.</p><p>In the spreadsheet, add habits in rows and dates in columns. Fill all the cells in those rows and columns with checkboxes to tick off daily progress. Then, use conditional formatting to make it look visually clear. Keep it simple: use green for success and amber for missed days.</p><p>Now for the measurement part: use the SUM formula at the end of the week or month to calculate your percentage of competence. All these features, without any nagging notifications and recurring payments, make Excel a tier-1 suggestion.</p><h3>4. Goal Setting — Break Big Goals Into Weekly Actions</h3><p>Your goal can be anything. It can be self-improvement or preparation for an exam. Regardless, one thing is always applicable: a goal without a structured system is just a wish.</p><p>What is that system?</p><p>Excel is a potential option. It breaks big annual goals into quarterly milestones, which further get divided into three to five weekly actionable tasks. You can easily use a customizable progress bar by using formulas and conditional formatting. For example, make the bar turn green as you close in on your target.</p><blockquote><strong><em>Tip:</em></strong><em> A habit of a weekly review will make the journey towards your goals more than easy.</em></blockquote><h3>5. Time Tracking — Find Out Where Your Hours Actually Go</h3><p>In a time management course, I saw an Excel sheet devised to record where an individual spends time in order to make us aware of how we lose our time. So, Excel for time tracking is not new.</p><p>It can help you meet your reality, like how much focused work you are actually doing. We often assume that we put in a quarter of a day into work, but actually, it was one-sixth of the day due to meetings and distractions.</p><p><strong>Workflow</strong> — When you sit down to work, log your start time, end time, category, and a brief description. Excel formulas will calculate the duration automatically.<br> Then, use <strong>Pivot Tables</strong> to generate a visual breakdown of where your hours went (deep work vs. administration).</p><h3>6. Data Cleaning and Automation — Stop Doing the Same Task Twice</h3><p>If your job is to deal with data analytics and dashboards, then you are probably aware of Excel. You can take it to the next level by automating manual reformatting and streamlining the same recurring weekly report.</p><p>You just need thirty minutes of time to set up a <strong>Power Query</strong>, which will record your planning steps. Once it is recorded, future reports can then be updated automatically with a single click.</p><p>Believe me, it will save you a ton of time, cut down manual labor, and leave you more free to play with data.</p><h3>7. Personal Dashboard</h3><p>A dashboard for all aspects of your life. It would be like a cockpit — a single master sheet that pulls data from all your other tracking sheets automatically. Excel makes this possible using linked cell references and basic charts.</p><p>It will give you a glanceable, real-time summary of your finances, habits, tasks, and goals without any manual copy-pasting. And, at one single place.</p><h3>Conclusion</h3><p>Options for productivity apps are aplenty, but juggling multiple apps feels like a task. It would be so easy if everything could be placed under one roof like a supermarket. Excel is a good option. It is worthy of a try.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=091c90913d3f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How I Cleaned a 5,000-Row Messy Excel File in 10 Minutes]]></title>
            <link>https://medium.com/@skillsatwork.skills/how-i-cleaned-a-5-000-row-messy-excel-file-in-10-minutes-9eff41541ccd?source=rss-443824bd1245------2</link>
            <guid isPermaLink="false">https://medium.com/p/9eff41541ccd</guid>
            <category><![CDATA[productivity]]></category>
            <category><![CDATA[data-cleaning]]></category>
            <category><![CDATA[excel]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[data-analysis]]></category>
            <dc:creator><![CDATA[Muhib]]></dc:creator>
            <pubDate>Sun, 29 Mar 2026 06:35:09 GMT</pubDate>
            <atom:updated>2026-05-14T08:53:04.718Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9IyiNjuhgD_ucFp15ZbB8A.jpeg" /><figcaption>Image generated by AI</figcaption></figure><p><em>I used to spend hours on this. Then everything changed.</em></p><p>When I was new to Excel, I would frown at data cleaning.</p><p>Every messy file meant the same thing — go through each row, manually correct each entry, one by one. The larger the data, the longer it took.</p><p>Tedious. Exhausting. Frustrating.</p><p>Then, one fine morning, I came across the TRIM() function.</p><p>I used it. It felt like a breath of fresh air.</p><p>That was the turning point. I started exploring. I added function after function to my arsenal. Tools that made data cleaning feel less like punishment and more like a skill. Not only was it fun — it made the whole process efficient and seamless.</p><p>Today, a 5,000-row messy file doesn&#39;t intimidate me. It takes me about ten minutes.</p><p>Let me show you exactly how.</p><h3>Before Touching Any Data — The Setup</h3><p>I learned this the hard way.</p><p>The first time I cleaned a large dataset, I made a mistake midway and had no way to go back. The original data was gone. I had to download the file again.</p><p>Now, before I touch a single cell, I spend two minutes on setup.</p><p>I split the workbook into three sheets:</p><ul><li><strong>Raw Data</strong> — the original file, untouched. I never edit this sheet.</li><li><strong>Cleaning in Progress</strong> — where all the work happens.</li><li>Cleaned Data — the final output, ready for analysis.</li></ul><p>Not always but when I have to study that data, I add a <strong>Unique ID column</strong> — a simple column of consecutive numbers (1, 2, 3...) in column A. This lets me restore the original row order at any point, no matter how much sorting or filtering I do along the way.</p><p>It makes the workflow hassle-free.</p><p>Two minutes of setup. Zero panic later.</p><p><strong>Tip</strong> — Never clean your original data. Always work on a copy and know how to get back to where you started.</p><h3>The File That Landed in My Inbox</h3><p>It was a 5,000-row CRM export. The sender’s message was short.</p><p>&quot;Can you make sense of this?&quot;</p><p>I opened it. That familiar sinking feeling came as I was looking at an absolute mess. It had —</p><ul><li><strong>Extra spaces</strong> hiding inside and around names and addresses</li><li><strong>Invisible characters</strong> from the export that were silently breaking formulas</li><li><strong>Inconsistent capitalisation</strong> — &quot;SARAH JOHNSON&quot;, &quot;sarah johnson&quot;, &quot;Sarah Johnson&quot; all in the same column</li><li><strong>Duplicate rows</strong> — the same customer appearing two or three times</li><li>No table structure — just a flat, unformatted range of cells</li></ul><p>Any messy data has these defects in different permutations &amp; combinations.</p><p>Five problems staring at me. Luckily, I have a clear sequence to fix them.</p><p>Let&#39;s go through each one.</p><h3>Step 1 — The One Formula That Does Three Jobs</h3><p><em>TRIM + CLEAN + PROPER</em></p><p>When I was starting out, I used these functions one at a time.</p><p>TRIM first. Then CLEAN. Then PROPER.</p><p>Eventually I realised — why not nest them all into one formula?</p><pre>=PROPER(TRIM(CLEAN(A2)))</pre><p>One stone, three birds.</p><p>This single formula fixes all three text problems at once. Here is what each layer does:</p><ul><li><strong>CLEAN</strong> — strips out invisible, non-printable characters. The ones you cannot see but that cause your VLOOKUP to return #N/A for no obvious reason. CRM exports are full of these.</li><li><strong>TRIM</strong> — removes all leading and trailing spaces. Interestingly, it collapses multiple spaces between words into one. &quot; Sarah Johnson &quot; becomes &quot;Sarah Johnson&quot;.</li><li><strong>PROPER</strong> — fixes capitalisation. &quot;SARAH JOHNSON&quot; and &quot;sarah johnson&quot; both become &quot;Sarah Johnson&quot;. There is a caveat associated with this function.</li></ul><p>I applied this formula across every text column in a helper column beside the original data.</p><p>Then — and this step matters — I copied the results and used <strong>Paste Special → Values</strong> to replace the formulas with the actual cleaned text. This locks the data in place and keeps the workbook fast.</p><h4>Why Paste special</h4><p>If you delete the original column after applying the function then helper column shows error. It because the reference for the formula is deleted.</p><p>So, after finishing the cleaning, copy the data &amp; use Paste Special → Values in the same helper column itself. It paste the values devoid of the formula.</p><p>Now delete the original column. Your cleaned data stays put.</p><h4>Bonus tip for web-imported data:</h4><p>Sometimes TRIM still leaves behind stubborn spaces. These are non-breaking spaces — a special character TRIM cannot touch. Before running the nested formula, use this:</p><pre>=SUBSTITUTE(A2, CHAR(160), &quot; &quot;)</pre><p>This replaces non-breaking spaces with regular ones. Then TRIM can finish the job.</p><p>Aside TRIM(), PROPER() makes the first letter of all the words capital. So words like USB, HDMI which needs to remain in block letter become Usb &amp; Hdmi.</p><p><strong>Fix — </strong>Use Find &amp; Replace to correct them.</p><h3>Step 2 — Split Columns in Seconds With Flash Fill</h3><blockquote>Flash Fill — Ctrl + E</blockquote><p>The file had full names in one column. &quot;Sarah Johnson&quot;. &quot;Mohammed Al-Rashid&quot;. &quot;Priya Sharma&quot;. I needed them split into First Name and Last Name.</p><p>The best of all the available methds is Flash Fill. It will change everything. It is a magical function.</p><p>I created a new &quot;First Name&quot; column. In the first cell I typed &quot;Sarah&quot;. In the second I typed &quot;Mohammed&quot;. By the third cell, Excel had already figured out the pattern — a ghosted preview of all 5,000 rows appeared.</p><p>I pressed <strong>Ctrl + E</strong>. Done. All 5,000 rows filled instantly. Correctly.</p><p>I repeated the same process for Last Name. Two minutes, zero formulas.</p><p>Flash Fill works for almost any pattern — splitting names, extracting email domains, reformatting phone numbers. If you can show Excel the pattern twice, it handles the rest.</p><p><strong>Tip — </strong>Make sure to check any row randomly to monitor the flash fill results &amp; avoid any descrepancies.</p><p>One thing to know: Flash Fill results are <strong>static</strong>. They do not update if the original data changes. That is fine for data cleaning — you want locked, stable output anyway.</p><h3>Step 3 — Find Duplicates Before You Delete Them</h3><blockquote>Conditional Formatting + Remove Duplicates</blockquote><p>I never delete duplicates blindly. I look at them first.</p><p>I selected the email column and went to <strong>Home → Conditional Formatting → Highlight Cell Rules → Duplicate Values</strong>. Every duplicate email turned red instantly.</p><p>A quick scroll confirmed — these were genuine duplicate records. Same customers, entered multiple times.</p><p>Only then did I go to <strong>Data → Remove Duplicates</strong>, selecting customer ID and email as my unique identifiers.</p><p>One click. 312 duplicate rows removed. 4,688 unique records remained.</p><p>Three things I always remind myself before running this:</p><ul><li>Run TRIM and CLEAN first. &quot;Sarah Johnson &quot; and &quot;Sarah Johnson&quot; look the same to a human but Excel treats them as different values if the trailing space is still there.</li><li>Choose your duplicate columns carefully. Checking every column means two rows are only flagged if they are identical in every single field.</li><li>The built-in tool catches exact duplicates only. Fuzzy duplicates — same person, slightly different spelling — need a manual eye.</li></ul><h3>Step 4 — Turn the Data Into a Proper System</h3><blockquote>Excel Tables — Ctrl + T</blockquote><p>This is the step I skipped for years. I thought it was just about aesthetics — banded rows and a blue header.</p><p>I was wrong.</p><p>With 4,688 clean, deduplicated rows ready, I selected the dataset and pressed <strong>Ctrl + T</strong>.</p><p>That one action turned a cleaned range into a living, structured dataset:</p><ul><li>Filters appeared on every column header automatically</li><li>New rows added at the bottom would inherit all formatting and formulas on their own</li><li>Formula references became readable —</li></ul><pre>=SUM(CustomerData[Revenue]) instead of =SUM(D2:D4689)</pre><ul><li>Sorting always moves the entire row together — no more accidentally scrambling one column while the rest stays behind</li><li>Any Pivot Tables or charts connected to this table update automatically when new data arrives</li></ul><p>The file that arrived as an unstructured export was now something I could actually work with. Besides, I can hand it to someone else without a lengthy explanation.</p><p>Cleaned data is now analysis-ready.</p><h3>Before and After</h3><p><strong>Before</strong>: 5,000 rows of raw CRM export. Spaces and invisible characters breaking formulas. Inconsistent capitalisation throughout. 312 duplicate records inflating every count and total. Full names in one column. No filters. No structure.</p><p><strong>After</strong>: 4,688 rows of clean, standardised, deduplicated data. Separate first and last name columns. Consistent capitalisation across every text field. An Excel Table with filters, expandable rows, and readable formulas. A file ready for analysis.</p><p><strong>Time taken</strong>: less than 10 minutes.</p><blockquote>Cleaning data manually is a tiring. The knowledge of which tools to reach for and in what order, makes it fun and quick.</blockquote><h3>What I Wish Someone Had Told Me Earlier</h3><p>When I was going through rows manually, I thought that was just how it worked. That data cleaning was supposed to be painful.</p><p>It is not.</p><p>Most messy datasets have the same five or six problems in different combinations. Invisible characters. Extra spaces. Inconsistent capitalisation. Duplicates. Unstructured columns. No table format.</p><p>Once you have a reliable sequence for each of those — and the muscle memory to run it without thinking — the cleaning phase stops being a day&#39;s work and starts being a ten-minute routine.</p><p>I wish I had known this earlier. I am writing it here so you do not have to figure it out the long way.</p><blockquote>Excel was meant to ease our tasks. It continues to do so.</blockquote><h3>The Full Sequence — Save This</h3><ol><li><strong>Set up your workbook</strong> — three sheets (Raw / Working / Clean) and a Unique ID column before you touch anything</li><li><strong>Deep clean your text</strong> — <em>=PROPER(TRIM(CLEAN(A2))) </em>across all text columns.</li><li><strong>Paste as Values</strong> — replace all formulas with their results before moving on</li><li><strong>Flash Fill (Ctrl + E)</strong> — split or restructure columns by showing the pattern twice</li><li><strong>Preview duplicates</strong> — Conditional Formatting → Highlight Duplicate Values before deleting anything</li><li><strong>Remove Duplicates</strong> — Data → Remove Duplicates, selecting your unique identifier columns</li><li><strong>Convert to Table (Ctrl + T)</strong> — lock everything into a structured, filterable, auto-expanding dataset</li></ol><p>Save this sequence. The next messy file is coming.</p><p>Now you are ready for it.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=9eff41541ccd" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>