<?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 Vinayak Rastogi on Medium]]></title>
        <description><![CDATA[Stories by Vinayak Rastogi on Medium]]></description>
        <link>https://medium.com/@rastogivinayak2004?source=rss-f86ba3f0ef54------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*MPVbArfQxl_O1b35ohscfw.jpeg</url>
            <title>Stories by Vinayak Rastogi on Medium</title>
            <link>https://medium.com/@rastogivinayak2004?source=rss-f86ba3f0ef54------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 02:43:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rastogivinayak2004/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[AI Made Me Question Why I Attend College. Then I Met This Professor…]]></title>
            <link>https://medium.com/@rastogivinayak2004/ai-made-me-question-why-i-attend-college-then-i-met-this-professor-2c5055791410?source=rss-f86ba3f0ef54------2</link>
            <guid isPermaLink="false">https://medium.com/p/2c5055791410</guid>
            <category><![CDATA[teachers]]></category>
            <category><![CDATA[students]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[classroom]]></category>
            <category><![CDATA[learning]]></category>
            <dc:creator><![CDATA[Vinayak Rastogi]]></dc:creator>
            <pubDate>Thu, 29 Oct 2026 17:27:14 GMT</pubDate>
            <atom:updated>2026-10-29T17:27:14.703Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sKHIAY4-gn0s0f5m0p0syA.png" /></figure><p><em>When assignments are created by AI, completed by AI, and graded by AI, what exactly are we going to classrooms for?</em></p><p>How long ago did you actually look forward to going to class? Not because you had to go, not because you had to take a test, but just because you wanted to learn something.</p><p>For a long time, I couldn’t remember feeling that way.</p><p>We live in an age where AI has become an integral part of everything in education. Professors employ AI to make assignments, students use AI to do assignments, and sometimes even those very assignments are marked by AI.</p><p><strong>An assignment created by AI, solved by AI, and graded by AI. Somewhere in between, there’s a student who may not have learned anything at all.</strong></p><p>But I’m not criticizing anyone who uses AI. I use it extensively myself. It’s probably one of the best learning aids available to us today.But it made me question something.</p><p>Now that I have an option of asking AI to explain a concept in five different ways, generate examples, and resolve my queries right away, why would I need to waste my time sitting in a classroom and listening to someone reading PowerPoint slides?</p><p>Gradually, the whole idea of learning at school became less exciting and more monotonous.</p><p>Attend lectures. Maintain attendance. Submit assignments. Prepare for exams. Repeat.</p><p><strong>Education had started to seem like an unending cycle of deadlines and paperwork, not an exchange of ideas.</strong></p><p>And honestly, I was almost convinced that this was simply how college worked now.</p><p>Until this semester.</p><h3>The Professor Who Changed My Perspective</h3><p>I am now studying in my penultimate year in college, and during this term, I had a teacher for Python Programming who totally transformed my perception of classroom learning.</p><p>And not because she employed any revolutionary style of teaching, but because she brought back what a good classroom should have felt like to me.</p><p>It was not simply a case of her entering the class, opening the presentation, going through the slides, and leaving the classroom.</p><p>She actually made sure we were learning.</p><p>Her classes were interactive. She encouraged discussions, asked questions, and regularly called students to the board to solve problems.</p><p>But what stood out to me the most was how she treated mistakes.</p><p>Usually, when students are called to the board, there’s always that little fear of getting something wrong in front of everyone. Even when you have an idea, you might hesitate because you don’t want to embarrass yourself.</p><p>But in her classes, making a mistake wasn’t something to feel guilty or embarrassed about.</p><p><strong>She appreciated students for trying, even when their answers were wrong.</strong></p><p>And that small gesture made a huge difference.</p><p>Students could attempt problems without worrying so much about being judged. The classroom felt like a place where mistakes were part of learning, not something to be ashamed of.</p><p>There wasn’t one particular incident that made her classes special. It was the way she consistently made even ordinary lessons enjoyable.</p><h3>She Didn’t Tell Us to Stop Using AI</h3><p>Another thing I genuinely appreciated was her approach toward AI.</p><p>While many educators are still struggling with whether students should be allowed to use AI, she didn’t simply ask us to stop using it.</p><p>Instead, she taught us how to use it.</p><p>She combined traditional, interactive classroom teaching with modern technologies, showing that AI doesn’t necessarily have to compete with teachers.</p><p>And that made me realize something.</p><p><strong>The future of education isn’t about replacing teachers with AI. It’s about having teachers who know how to teach in a world where AI exists.</strong></p><p>AI can explain Python syntax, generate code, and help debug errors. But having a classroom where you’re encouraged to think, solve problems, make mistakes, and learn alongside others is a different experience.</p><p>And I had almost forgotten how valuable that could be.</p><h3>For the First Time, I Actually Wanted to Prepare for Class</h3><p>Here’s the part that surprised me the most.</p><p>I started revising previous lectures before attending the next class.</p><p>Not because there was an assignment due, or because I was worried about my grades.</p><p>I simply wanted to keep up with what she was teaching.</p><p>Sometimes, I’d spend far more time preparing for small quizzes or assessments than was actually necessary.</p><p>And nobody was forcing me to do any of it.</p><p>I wanted to.</p><p>For the first time in a long while, studying didn’t feel like something I had to do just to meet academic requirements.</p><p>It became something I genuinely enjoyed.</p><p>And there was another reason I found myself putting in more effort.</p><p>I didn’t want to disappoint her.</p><p>Not because she was strict or intimidating, but because I could see how much effort she put into making every class meaningful.</p><p>When someone is trying so hard to make an entire classroom learn, you naturally feel like putting in some effort yourself.</p><p>I often thought that if she was putting so much energy into teaching us, the least I could do was come prepared and genuinely try.</p><p>That motivation didn’t come from fear.</p><p>It came from respect.</p><p><strong>Sometimes, the best motivation to study isn’t grades or competition. It’s having a teacher whose efforts make you want to give your best.</strong></p><h3>Maybe AI Isn’t the Problem After All</h3><p>Before this semester, I was beginning to believe that AI had made traditional classroom learning somewhat unnecessary.</p><p>But now I think the problem isn’t necessarily AI.</p><p>Maybe students aren’t losing interest in classrooms because AI can explain things better.</p><p>Maybe they’re losing interest because too many classrooms have become passive experiences.</p><p>A classroom shouldn’t be a place where students spend an hour copying information they could have accessed in seconds.</p><p>And a teacher’s role shouldn’t be limited to delivering information that’s already available everywhere.</p><p>A good teacher creates curiosity, encourages questions, recognizes effort, and makes students comfortable enough to be wrong.</p><p>AI can provide explanations, examples, and solutions whenever we need them.</p><p>But a great teacher can make us <em>want</em> to understand those explanations in the first place.</p><p><strong>AI can make information easier to access. A great teacher can make you care enough to seek it.</strong></p><p>And I think that’s the difference I had been missing.</p><h3>So, Is Classroom Learning Dead?</h3><p>I don’t think so anymore.</p><p>But I do believe classrooms need to evolve.</p><p>We can’t expect students who have access to powerful AI tools and unlimited learning resources to remain excited about lectures that offer nothing beyond a one-way transfer of information.</p><p>The role of a teacher is changing, and perhaps that’s a good thing.</p><p>Because the most valuable teachers were never valuable simply because they knew more than their students.</p><p>They were valuable because they knew how to make students learn.</p><p>This semester reminded me that classrooms can still be meaningful, engaging, and genuinely worth attending.</p><p>I still use AI to learn, and I probably always will.</p><p>But now I also know that there are classes I genuinely look forward to.</p><p>Not for attendance.</p><p>Not for grades.</p><p>Not for deadlines.</p><p>But simply because I enjoy learning there.</p><p>And perhaps the most important thing I learned this semester wasn’t even part of the Python syllabus.</p><p><strong>A great teacher doesn’t just make a subject easier to understand. They make you want to understand it.</strong></p><h3>A Small Thank You</h3><p>To my Python professor, if you ever happen to read this, thank you.</p><p>Thank you for making learning enjoyable again, for appreciating attempts even when answers were wrong, and for creating a classroom where students could participate without being afraid of making mistakes.</p><p>You probably don’t realize how much of a difference your teaching approach has made.</p><p>But at least for one student, you changed the way he looks at classroom learning.</p><p>And that’s something I’ll remember long after this semester ends.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2c5055791410" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Debugging Taught Me More Than AI Ever Could]]></title>
            <link>https://medium.com/@rastogivinayak2004/debugging-taught-me-more-than-ai-ever-could-0dd5eea9415f?source=rss-f86ba3f0ef54------2</link>
            <guid isPermaLink="false">https://medium.com/p/0dd5eea9415f</guid>
            <category><![CDATA[aritificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[technology-and-society]]></category>
            <dc:creator><![CDATA[Vinayak Rastogi]]></dc:creator>
            <pubDate>Mon, 10 Nov 2025 05:09:38 GMT</pubDate>
            <atom:updated>2025-11-10T05:09:38.571Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="cover-image" src="https://cdn-images-1.medium.com/max/670/1*oWQ4wk9ZRzuokifsHVvrmw.jpeg" /></figure><p>When I started coding, the process was slow, messy, and incredibly satisfying. I wasn’t just learning a language — I was learning how to think. Every red error message on the screen forced me to pause, trace my steps, and understand <em>why</em> something went wrong. Debugging was frustrating, yes, but it was also my teacher.</p><p>Now, years later in college, I see something different. The new generation of learners, many of whom are my friends, rarely experience that same process. When their AI-generated code breaks, they don’t even read the error. They copy it, paste it into ChatGPT or another tool, and move on. The fix appears in seconds, but the understanding never does.</p><h3>The lost art of learning through failure</h3><p>Once upon a time, every error message was an invitation to explore. A missing semicolon taught you syntax. A segmentation fault taught you about memory. An exception made you understand control flow. You <em>earned</em> every line of code that worked.</p><p>That slow grind built patience, intuition, and curiosity. You didn’t just know <em>how</em> to fix something — you began to sense <em>why</em> it broke. Each bug left a faint scar that made you a bit sharper the next time.</p><p>Now, AI removes that friction. It corrects the problem instantly, but it also erases the learning curve that used to shape us. The pain was part of the process, and in removing the pain, we’ve lost the growth that came with it.</p><figure><img alt="terminal with errors" src="https://cdn-images-1.medium.com/max/992/1*p-fXSh2HwXFrIO7QEwxe1g.jpeg" /></figure><h3>Developing more than documenting</h3><p>We’re producing more code than ever before, but understanding less of it. In the rush to build faster, we’re documenting less, reflecting less, and asking fewer questions.</p><p>There was a time when writing code also meant explaining it — to yourself and to others. We left comments, wrote README files, or discussed design choices with teammates. Today, AI tools can generate all that too, so we skip the thinking part.</p><p>The result is a quiet kind of knowledge loss. We have working systems but fewer developers who truly understand their inner logic. Code is multiplying, but comprehension is thinning out.</p><h3>The disappearance of exploration</h3><p>Before AI, solving a problem meant diving into Stack Overflow, GitHub issues, or developer forums. That process wasn’t just about finding the right answer — it was about encountering <em>different minds</em>.</p><p>You might visit a thread to fix one specific bug and end up learning five new concepts. Someone would post one solution using method X, and another person would counter with method Y, explaining why it’s better in the long run. Even if you didn’t adopt Y, you left the page knowing something you didn’t know before.</p><p>Now, AI delivers one clean solution, stripped of that messy, human debate. It saves time, but it also filters out the learning that used to happen by accident — the discoveries we made while <em>searching</em>. We no longer stumble upon ideas; we just consume finished answers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9xVeAeOJQuHOkSJ452wyXg.png" /></figure><h3>AI as a mirror, not a mentor</h3><p>AI can reflect what we already know, but it can’t replace the thinking process that creates understanding. It gives us answers, not intuition. When we ask it to fix something, we often stop asking <em>why it needed fixing in the first place.</em></p><p>That shift is subtle but dangerous. We’re becoming dependent on tools that think for us, and in doing so, we risk losing our ability to think deeply at all. AI can assist, but it can’t struggle for us — and struggle was where the real learning happened.</p><h3>The next generation builds fast, but understands less</h3><p>Speed has become the new definition of skill. Students can now write entire programs in minutes, but when something breaks, many don’t know where to look. They can run code, but they can’t trace it. They can execute, but not explain.</p><p>This doesn’t mean they’re lazy — it means the ecosystem around them encourages results over reasoning. AI shortens the path between idea and execution, but it also shortens the space where thinking used to live.</p><p>It’s like taking a test with all the answers already filled in. You’ll pass easily, but you won’t understand the subject.</p><h3>A quiet warning</h3><p>AI isn’t the villain here. The real danger is what happens when we stop questioning. The less we debug, the less we learn how to think like programmers. We risk creating a generation of coders who can generate solutions but not solve problems.</p><p>Maybe debugging was slow. Maybe it was painful. But it was also where we learned persistence, logic, and respect for complexity. Those were not just programming skills — they were life skills.</p><p>We used to fix errors because we <em>had to</em>. Now AI fixes them for us because it <em>can</em>. But in handing over that small act of struggle, we might be giving away something much larger — the part of learning that makes it truly ours.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0dd5eea9415f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How multithreading can affect the efficiency of a process ?]]></title>
            <link>https://medium.com/@rastogivinayak2004/how-multithreading-can-affect-the-efficiency-of-a-process-03c5e913e6b8?source=rss-f86ba3f0ef54------2</link>
            <guid isPermaLink="false">https://medium.com/p/03c5e913e6b8</guid>
            <category><![CDATA[multithreading]]></category>
            <category><![CDATA[concurrent-programming]]></category>
            <category><![CDATA[asynchronous-programming]]></category>
            <category><![CDATA[threads]]></category>
            <category><![CDATA[programming]]></category>
            <dc:creator><![CDATA[Vinayak Rastogi]]></dc:creator>
            <pubDate>Sat, 28 Dec 2024 07:30:01 GMT</pubDate>
            <atom:updated>2024-12-28T07:30:01.239Z</atom:updated>
            <content:encoded><![CDATA[<h3>How multithreading can affect the efficiency of a process ?</h3><p>Before we start discussion how using threads in different ways in program affects its efficiency, we should first know how thread works and different ways to achieve multithreading.</p><h3>Threads</h3><p><strong>What is a Thread and why you should use it in your program?</strong></p><p>A set of instructions saved in computer’s secondary memory is called a Program and when this program is loaded in RAM for execution it’s called a process or simply “ <em>a process is a program under execution </em>”. Within these processes, threads exists as lightweight, independent sequence of execution.</p><p>Some people tend to confuse between a process and thread, basically “ <em>a process is an independent program with separate memory, while a thread is a smaller, concurrent unit within a process that shares memory </em>”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Q2g0JidnRcdWl8uFgyHvog.png" /><figcaption>Visualization of Program Execution Flow: From Disk to Processes and Threads in RAM</figcaption></figure><p>Using threads in a program can provide several benefits, especially when tasks can be performed concurrently or when your program needs to handle multiple operations at the same time. Using threads can result in Concurrency and Multitasking, Improved Performance, Responsive Applications, Parallelism, Simplified Design, Real-Time Processing, Efficient Resource Sharing, etc.</p><h3><strong>MultiThreading</strong></h3><p>It is a programming technique that allows a process to have multiple threads running concurrently, enabling tasks to execute simultaneously within the same program. This approach improves performance by allowing a program to take advantage of multiple CPU cores.</p><p><strong>Concurrency and Parallelism</strong></p><p>Concurrency and parallelism are general principles for structuring and executing tasks in a program. They describe how tasks are managed and executed.</p><p><strong>Concurrency</strong> is the ability of a program to manage multiple tasks at once. However, these tasks might not actually be running simultaneously. Instead, the system might switch between them (context switching) to give the illusion of simultaneous execution.</p><p><strong>Parallelism</strong> is the execution of multiple tasks at the exact same time. It requires multiple processors or cores, where each task is assigned to a separate core and runs simultaneously.</p><p><strong>Multithreading</strong> is a way to achieve concurrency or parallelism in a program under execution.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uDluabHwkKJy2G9poup_KQ.png" /><figcaption>Execution of Threads under Concurrent execution and Parallel execution</figcaption></figure><p><strong>How multithreading affects the efficiency of a program?</strong></p><p>To demonstrate this we are gonna take a program i.e. Matrix Multiplication on large scale (larger matrices with greater values), then we will calculate time taken by N number of threads (N ≥ 1 and N ≤ number of rows in matrix) to perform matrix multiplication and then compare them.</p><p><a href="https://github.com/vinayakrastogi/threads_performance">GitHub - vinayakrastogi/threads_performance: A single process (matrix multiplication on large matrices) is completed using multiple threads and time is noted to demonstrate how multithreading can affect performance.</a></p><p>We are going to implement it through parallelism i.e. by dividing the number of rows equally among the threads. For example if we have two matrices of size 2000x2000 and we will be using 50 threads, so each thread will be handling equal amount of operations i.e. 40 rows will be handled by each thread, this might not sound like it’s doing much but this process is about 8 billion multiplication operations and 7.996 billion addition operations performed by the CPU. The results will be stored in a log file and used to generate a line-graph.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/882/1*F5-mufZ09hkJ77Ey40LBvA.png" /></figure><p>Initially, the time decreases significantly as the number of threads increases, demonstrating the benefit of parallelism and efficient utilization of CPU cores. However, beyond a certain point, the graph exhibits fluctuations and spikes instead of a smooth decline. These spikes can be attributed to factors such as thread overhead, where the cost of managing and synchronizing threads becomes significant. Hardware resource contention, including competition for CPU cores, cache memory, and memory bandwidth, also plays a role, especially as the number of threads exceeds the physical cores available. Additionally, operating system thread scheduling and imbalanced workload distribution can lead to inefficiencies, causing abrupt increases in execution time. The final spike near the maximum thread count may reflect severe contention or diminishing returns from adding more threads, as the overhead outweighs the parallelism benefits.</p><p>Generally, increasing the number of threads doesn’t always lead to faster execution, as it depends on both the process and the system. Processes with high synchronization requirements or shared resource contention can experience diminishing returns or slower performance with more threads. Similarly, the system’s hardware, such as the number of cores and memory bandwidth, plays a crucial role in determining the optimal number of threads for efficient execution.</p><h4><strong>Industrial Use</strong></h4><p><strong>Parallel execution</strong>, is essential in scenarios demanding high computational power and efficiency. It is widely used in high-performance computing (HPC) tasks such as weather simulations, molecular modeling, and scientific research. Additionally, fields like artificial intelligence (AI) and machine learning rely heavily on parallelism, where GPUs process large datasets and train models by performing many computations simultaneously, significantly reducing the time required for these tasks.</p><p><strong>Concurrent execution</strong>, on the other hand, is commonly used in areas where responsiveness and multitasking are crucial. For example, web servers handle multiple client requests concurrently, ensuring that each user’s request, such as loading a webpage or submitting a form, is processed without delay. Similarly, mobile applications like messaging apps use concurrency to allow users to send messages, receive notifications, and sync data in the background, creating a seamless user experience.</p><p><strong>Hybrid use</strong>, Modern systems often combine concurrency and parallelism to achieve both responsiveness and performance. For instance, web browsers handle multiple tabs and user interactions concurrently, ensuring smooth responsiveness, while using parallelism to accelerate tasks like rendering webpages or processing images using multiple cores or GPUs. Similarly, cloud computing platforms use concurrency to manage thousands of user requests simultaneously, while employing parallel processing to handle computationally intensive tasks, such as analyzing big data or training machine learning models. This hybrid approach leverages the strengths of both techniques, enabling systems to handle complex workloads efficiently while maintaining high interactivity.</p><p>In <strong>conclusion</strong>, both concurrent and parallel execution are fundamental to building efficient and scalable software systems. While concurrency ensures responsiveness and better resource utilization, parallelism enables faster execution of computationally demanding tasks. Together, they form the backbone of modern applications, from web servers to artificial intelligence and cloud computing. Understanding their strengths, limitations, and optimal usage is essential for developers and engineers to design systems that meet the demands of today’s technology-driven world.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=03c5e913e6b8" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>