Outlier’s Post

In my previous post, I discussed the concerns raised by one of my students about the rapid advancements in AI and its potential impact on the future of software development. Now, I'd like to address the value of introductory computer science education in the face of these changes. Firstly, it's important to recognize that becoming a software engineer is not the most common outcome for students in my IB Computer Science course. Instead, I believe that the long-term personal utility gained from this class, or any introductory CS exposure, comes from shifting a student's mental model of how modern social and economic systems interact with and depend on these technologies. While the cutting-edge AI tools we see today are far more advanced than the simple Von Neumann CPU models and introductory Python we cover in class, our curriculum still addresses the foundational concepts that underpin these systems. Similarly, our Higher Level Analysis and Physics courses may not delve into the specific mathematics and physics behind large-scale machine learning models, but they provide the necessary groundwork for understanding these complex systems. If a student's concern is, "The world is changing rapidly, this class is challenging, and I don't think I'll find a 2022-style entry-level software development job when I graduate from university in 2029," I would wholeheartedly agree with that sentiment. However, this raises a broader question: How do we plan for the unknown? Schools are inherently backward-looking to some degree, as they rely on established knowledge and practices. Yet, we must consider what aspects of our current education system, combined with insights from the past, might offer the best expected value for navigating the uncharted territory that lies ahead. #Education #FutureOfWork #ArtificialIntelligence #AdaptiveLearning

I'm happy to announce that I defended my dissertation last Friday at Aalto University. Short conclusion: Novices are not better at answering questions about code, that they have written and works correctly, than about example codes. Students who frequently answer questions about their code incorrectly are prone to tinkering with their code and seemingly lack design while they work. Those results predate popular LLMs, which can generate programs and answer questions about them better than average novices, although their answers included many humanlike errors. Automated questions about students' code could offer practice in program comprehension, practice towards discussing program code, support at the level of metaknowledge, and early warnings about students who need help although they produce programs. In the era of AI-assisted programming, communicating the importance of reading, reviewing, and discussing code is important, as well as offering students tasks to evaluate LLM outputs and thus calibrate their trust.

"Hello, LinkedIn! 👋 I’m excited to join this professional platform as a third-semester Computer Science student. I’m eager to connect with professionals, share insights, and learn from others in the tech field. Currently, I’m diving into Javascript or more and more areas of CS such as AI and Software development. Feel free to connect, and let’s share knowledge and ideas! #ComputerScience #TechStudent #Learning #Networking #StudentJourney"

“It's hard to give advice. At least as hard as to take advice. Don't over-specialize, don't be too sure that you know the future, be flexible, and remember that careers and jobs are a long-term thing.  Too many young people think they can optimize something, and then they find that they’ve spent a couple of years, or more, specializing in something that may not have been the right thing, and in the process, they burn out. Because they haven't spent enough time building up friendships and having a life outside computing.  I meet a lot of sort of.. I don't know what we can call them, junior geeks that just think that the only thing that matters is their sub-specialty of computing, programming, or AI, or graphics or something like that. And well, it isn't. And the rug might be pulled under them but for that and if they do nothing else.. Well, if you don’t communicate your ideas, you could just as well do Sudoku. You have to communicate and a lot of a lot of sort of caricature nerds forget that.  They think that if they can just write the best code, they change the world, but you have to be able to listen, you have to be able to communicate with your would-be users and learn from them, and you have to be able to communicate your ideas to them. So you can't just do code, you have to do something about culture and how to express ideas, and I mean, I never regretted the time I spent on history and on math. Math sharpens your mind, history gives you some idea of your limitations on what's going on in the world, and so don't be too sure to take time for to have a balanced life and be ready for the opportunity, I mean a broad-based education, a broad-based skill set, which is what you build up when you educate, you're basically building a portfolio of skills, means that you can take advantage of an opportunity when it comes along. You can recognize it sometimes.  We have lots of opportunities but a lot of them we either can't take advantage of or we don't notice.   It was my fairly broad education: I've done standard computer science, I've done compilers,  I've done multiple languages, I think I knew two dozen at the time, and I have done machine architecture, I've done operating systems,  and that skillset turned out to be useful.” Bjarne Stroustrup (73yo), the Creator of C++ You can view the full interview on YouTube: https://lnkd.in/dVXQraX6

Embarking on a 251-Question DSA Journey 🚀 Hello, LinkedIn community! I'm excited to share that I've decided to take on a new challenge to deepen my understanding of Data Structures and Algorithms (DSA). Over the next few months, I will solve 251 DSA questions curated by Parikh Jain, and I'll share my daily progress here on LinkedIn. Why DSA? Data Structures and Algorithms are the backbone of computer science. Mastery of these concepts is crucial for roles in software development, backend development, frontend development, full-stack web development, and even data science. These questions are designed to cover the most common problems asked in technical interviews across various companies, from startups to major product-based firms. The Roadmap: Here's a brief overview of the roadmap I'll be following: Overall Basic DSA Progress: - Total Questions: 50 Overall Intermediate DSA Progress: - Total Questions: 75 Overall Advanced DSA Progress: - Total Questions: 100 DSA Project: - Huffman Coding Miscellaneous DSA: - Total Questions: 25 Follow Along! I'll be posting daily updates, including: - The problem I tackled - The challenges I faced - The solution approach - Key learnings and insights This journey will enhance my problem-solving skills and keep me accountable and motivated. If you're interested in DSA or looking to improve your coding skills, feel free to follow along or join me in this challenge! Here's a useful video:https://lnkd.in/gcxPec_h explaining how to use the roadmap effectively. Let's embark on this learning journey together! #DSA #CodingJourney #Learning #DailyProgress #Tech

I self-study most of my life. Learned to read when I was 2: parents reasonably didn’t want to teach me yet, but I looked over how my brother does it and figured it out. Learned chemistry basics using my brothers’ textbooks - this gave me a boost when we started to learn in my class. Then I got interested in software engineering, while studying electrical engineering. In the good old days we had no video lectures, not even good internet access, and books were the only reasonable medium. Some books were very extremely helpful, allowing me to get into a new programming language, area of mathematics, or new engineering paradigm, others - not so much. There could be a great book that my friends used in their university and I couldn’t get through it. And sometimes I would discover some less known esoteric book and have a great adventure working through it. Now I have a hypothesis about what makes a good book for self-study. They provide chains of thought connecting the flow through the formulas, definitions, and proofs. They are much more verbose in a good way. “What if we do it this way? No, this is not gonna work because of that. Then can we do it this way? Also not. But how about this - oh, this is helping.” Seeing how the internal thought process is important for AI systems as o1 makes this even more glaring. For example, ~10 years ago I was studying in the MIT Linear Algebra class led by Gilbert Strang. The class has a set of video lectures, recommended book chapters (by the same author), homework, and quizzes. I tried to read some parts of the books and could not get past the first few pages. But once I watched the lectures on the same subject, the missing connecting pieces were in my head and I could be much further. The lectures provided intuition behind the new ideas, presented the development of the solutions step-by-step, with visualisations and the very detailed author comments. Another book that was hard to get through is the Elements of Programming by Alexander A. Stepanov (the father of C++ STL). I watched an amazing lecture with him about the book and got captured. But the book was very formal, rigorous, and dry; how the author believed the good technical books should be written. A good counter-example? Lectures of Feynman. Even in written form they provide different ways of thinking about how the world around us works. Feynman can explain the most complicated phenomena using simple abstractions and build most of the formulas from first principles. That’s also why some video lectures (Salman Khan, 3blue1brown) are so helpful for many students. They mostly consist of chains of thoughts of extremely talented teachers. P.S. I love how building artificial intelligence helps human better understand how their own intelligence works. P.P.S. The availability of chains and tricks of thoughts is also probably why some professions are quite often spread in families.

🚀 Over the past month, I've seen several posts detailing what a CS student should do to get ahead of the game. Most of these posts mentioned committing a lot on GitHub. 🤔 Here's the thing: CS is computer science, not software engineering. According to Merriam-Webster, computer science is a branch of science that deals with the theory of computation or the design of computers. 💡 In other words, computer science is more about the fundamental principles that make computers work and the foundation of software development. It delves into topics like algorithms, data structures, computational theory, machine learning, artificial intelligence, and more. These concepts are crucial as they form the basis upon which software applications and systems are built. 👀 Yes, committing on GitHub can be crucial as it can be shown as proof that you were part of a project, but that's not the only way. Recruiters look at your experience and see how much you have learned and grown rather than just the commits and how much you contribute to a project daily. 🎯 Here are things you can focus on this summer: 📘 Deepen Your Understanding of Core Concepts: Instead of just coding, spend time understanding the theoretical underpinnings of computer science. This will give you a solid foundation that can be applied in various contexts. 🔍 Engage in Research: Participate in research projects with professors or research groups. This can provide you with a deeper understanding of theoretical concepts and expose you to cutting-edge developments in the field. 🌐 Explore Interdisciplinary Applications: Look into how computer science intersects with other fields like biology, physics, or economics. Understanding these connections can open up new perspectives and opportunities. 🤝 Develop Soft Skills: Communication, teamwork, and problem-solving skills are crucial in any career. Participate in group activities, public speaking events, or leadership roles to hone these skills. 💼 Gain Practical Experience: Internships and co-op programs can provide hands-on experience and demonstrate your ability to apply theoretical knowledge in real-world settings. This experience is invaluable and often more impressive than just code contributions. 📈 Stay Current with Trends: The tech industry evolves rapidly. Stay updated with the latest trends, tools, and technologies by reading research papers, following tech news, and taking online courses. 🤝 Network with Professionals: Attend conferences, seminars, and meetups. Networking with industry professionals can open doors to new opportunities and provide insights into the industry's demands and expectations. 🌱 Remember, computer science is a vast field with numerous paths to explore. Focus on building a solid foundation in computer science principles and continuously learning and growing.

🚀 Excited to Share My Personal Portfolio! 🚀 I'm thrilled to unveil my personal portfolio, where I've compiled my journey as a B.Tech Computer Science student and my passion for AI, coding, and research. 🌟 Explore My Portfolio: https://lnkd.in/gqvxdx9J Inside, you'll discover: 🧠 Skills: Proficiency in Python, Java, C++, and more. 🔬 Experience: Insights from my time as an NSS Volunteer and Research Intern at the Technology Business Incubator. 💼 Projects: A showcase of my latest work, from AI-driven applications to software development. 🎓 Certifications: Including Azure AI Fundamentals and Python Programming. 📄 Download My Resume: Ready for new opportunities and collaborations! I'm constantly seeking to learn, grow, and contribute to innovative projects. Let's connect, exchange ideas, and work on something extraordinary together! 🌍 #Portfolio #AI #Coding #Research #TechInnovation #CareerGrowth #LetsConnect

🚀 Halfway through the Program Structure adventure trail! 🌟 As TA for this exhilarating journey, I've been privileged to witness the exploration of the foundational stones of Programs and Algorithms alongside Professor Siddhartha Nath. 📚 From the outset, Professor Siddhartha Nath crafted a roadmap guiding students through the various checkpoints along the trail. 🗺 So far, we've conquered five main checkpoints, each a triumph in its own right: 📍 Checkpoint-1: We delved into the fundamentals of asymptotic growth of functions and mastered comparison-based sorting algorithms like insertion sort and quicksort. 📍 Checkpoint-2: Our expedition led us to explore Python libraries like Numpy, where we honed our skills in data structures like stacks, queues, and LinkedLists. Plus, we navigated through the intricate pathways of sorting algorithms and recursion techniques. 📍 Checkpoint-3: Venturing into the realm of graphs and graph algorithms, we uncovered the secrets of basic representations and embarked on advanced traversal techniques, all while detecting elusive cycles along the way. 📍 Checkpoint-4: Our journey took a scenic route through the verdant landscapes of tree data structures, binary search trees, and traversals. With Python in hand, we braved the wilds of implementations, emerging triumphant. 📍 Checkpoint-5: Equipped with essential algorithms like Dijkstra's for navigating shortest paths and Kruskal's for charting minimum spanning trees, our toolkit is now fully stocked for the challenges ahead. In my role as TA, I've provided uninterrupted support to our students: 🔍 Offering guidance to ensure students grasp course materials. 💼 Assisting with Homework, mid-term and Programming Assignments, offering insights to help students excel in their coursework. 💬 Leading engaging after lecture discussions sessions, fostering interactive learning experiences that deepen understanding. 📝 Providing feedback on assignments, helping students assess their progress and refine their skills. As we stand at this midway point, the trail beckons with promises of further discovery and triumphs yet to be achieved. Here's to the next leg of the adventure! How is the Josh? 🌳💼 #AdventureTrail #ComputerScience #Algorithm Northeastern University #TeachingAssitant

In an era of AI and no-code tools, is learning to code still relevant? A Computer science student asked this question during a CSR interaction The short answer is: absolutely. While AI can automate many tasks, it's built on the foundation of code. Understanding programming languages allows you to: 1. Think critically: Coding teaches logical reasoning and problem-solving skills. 2 .Innovate: By learning to code, you can build your own solutions and ideas. 3. Collaborate effectively: Coding fosters teamwork and communication skills. 4. Future-proof your career: As technology evolves, coding skills will remain in high demand. By learning a programming language, you're not just acquiring a skill, you are developing a mindset. One of the key skills to accelerate your career in the tech industry, is your ability to learn, unlearn and relearn new technologies, skills etc. and learning a programming language helps to build a good foundation for same Do you agree that learning to code is important? Thanks Sankara Mahalingam for organizing the #CSR activity and providing young talented students an opportunity to visit the Firstsource campus and interact with the tech leadership team ! #coding #programming #computerscience #technology #futureofwork #AI #artificialintelligence