#29 — Staying relevant without getting overwhelmed
Follow signals that last
Learning CS with purpose
A video featuring four Spanish-speaking software engineers sharing how they’re navigating the AI wave in their day-to-day work. It offers a highly relevant perspective—both technically (as they are software engineers, architects, and developers) and from a business standpoint, with real-world use cases discussed. The conversation provides practical insights straight from the field:
The nature of the shift
If something can be computed, it will eventually fall into AI’s domain. The ceiling isn’t creativity or intent—it’s computability. This reality underscores a broader truth: programming has never been static. From assembly to high-level languages, from frameworks to natural language interfaces, constant evolution has always been the norm. What’s shifting now is not the existence of change, but its pace and surface area.
Yet the magic was never in the code itself—it’s in the business rules, the understanding of real problems, and the solutions we design. Code is an implementation detail. As tools evolve to handle more of that detail automatically, the software engineer’s value increasingly lies in clarity of thought, communication, and product insight.
What this means for Software Engineers
New market niches will emerge. Some jobs will disappear, but new roles and opportunities will open—often in areas we can’t fully anticipate yet. Software engineers are uniquely well positioned to lead this transition. It won’t be driven by veterinarians or lawyers; software engineers sit at the center of the shift.
Fundamentals still matter—especially when it comes to maintaining, debugging, and evolving software systems. AI can generate code quickly, but understanding why something breaks, how to fix it, and how to extend it still requires strong foundations. Senior engineers, equipped with both experience and AI tools, have become exponentially faster. Meanwhile, juniors won’t disappear—they’ll mutate. Knowledge of software engineering principles is becoming more important than mastering any specific language or framework.
Hussein Nasser also studies software engineering by focusing on fundamentals rather than trendy technologies. He starts by asking questions and digging deep. He gives an example involving networks, which illustrates his way of learning—starting from a specific doubt and exploring key related concepts. That’s when deep learning truly happens: when you investigate out of real need.
Interfaces, education, and learning
Despite predictions to the contrary, frontend development is not dead. Interfaces may evolve, but user habits, context, and human behavior still demand thoughtfully designed interaction layers. Similarly, how we teach programming matters more than ever. Education must balance foundational knowledge with modern tools—and crucially, preserve the value of struggle. Learning happens by making mistakes, debugging, and understanding why things fail. Removing that struggle entirely risks producing developers who can prompt but not problem-solve.
Being able to express what you’re thinking is now a critical technical skill. Communication—both with humans and with AI—has become central to effective software engineering. AI can either amplify you or shoot you in the foot, and the outcome depends entirely on the user’s understanding and intent.
Focus on what lasts
To stay current: study with purpose. Focus on what you need, ignore most hype, follow signals that last, read foundational papers, and be selective. Remember that LLMs are not all of AI—they’re powerful, but just one subset of a much broader field. Learn the basics well enough to guide AI tools intelligently. Your brain is the best resource you’ll have for the rest of your life—keeping it sharp is not optional anymore. Otherwise, you’re not engineering; you’re just reacting.
The best AI learning resources in 2026 span foundational books, research blogs, university courses, newsletters, and papers. Together, these resources provide both theoretical foundations and practical insights into modern AI systems.
An optimistic outlook
Soft skills are now non-optional. Communication, context-setting, and product thinking are fully part of the job. But this shift also brings opportunity: less meaningless labor, more leverage, more impact—and we’re only at the beginning.
The role of the software engineer is evolving—from assembly, to high-level languages, to natural language, and now toward orchestration. We’re moving from writing everything by hand to directing extremely powerful tools. The question is not whether this transition will happen, but how thoughtfully we navigate it.
Students and AI Literacy with Annette Vee
How do students really feel about generative AI and learning? What kind of guidance are they looking for from their instructors? And how can educators better understand students so they can adapt together to a world shaped by generative AI?
This week on the Intentional Teaching podcast, Annette Vee, Associate Professor of English at the University of Pittsburgh, joins Derek Bruff for a wide-ranging conversation about students, AI, and teaching.
Annette and her colleagues have spoken with hundreds of students at Pitt about AI, and she shares what she’s learned about how conflicted many students feel about AI’s role in higher education—and what it takes to have open, productive conversations about a topic that can be surprisingly difficult to discuss.
The conversation is part of an ongoing collaboration: Annette Vee and Derek Bruff are co-authors, along with Marc Watkins, on the forthcoming The Norton Guide to AI-Aware Teaching.
Like Marc, Annette was experimenting with LLMs before most educators had heard of ChatGPT. Her research sits at the intersection of computation and writing, and her first book, Coding Literacy: How Computer Programming Is Changing Writing (MIT Press, 2017), reflects that interdisciplinary background. Annette was directing the composition program at Pitt when ChatGPT launched in late 2022, and since then she has been helping colleagues at Pitt and beyond practice what she calls AI-aware teaching.
The episode covers a wide range of topics: how computational literacy is changing in light of AI, whether “AI literacy” is even the right term, what Annette has learned from listening closely to students over the past few years, and why AI needs a place in the college curriculum. Along the way, there’s also a brief detour into early programming experiences with BASIC and Logo in the 1980s.
You can listen to the conversation here:
🔍 Resources for Learning CS
→ MIT 6.S191: Introduction to Deep Learning
This fully open course is perfect for anyone looking to either dive into or refresh their deep learning knowledge.
→ The Concise TypeScript Book
An open-source book that offers a clear guide to TypeScript, suitable for both beginners and experienced devs.
→ Compilers and Interpreters Recourses
Here are some great options, organized by difficulty level:
Easy: Crafting Interpreters — Free and beginner-friendly
Medium:
Writing An Interpreter in Go — Paid
Writing A Compiler in Go — Paid
Advanced: Let’s Build a Compiler by Nora Sandler — Paid
→ Refactoring English
Refactoring English by Michael Lynch argues that developers aren’t bad writers—they just haven’t practiced deliberately. It shows how to apply programming-style best practices to write clearer docs, blogs, emails, and tutorials that developers actually read.
→ Data Science Resources
A website curated by Nicola Rennie that brings together free data science resources with filters by resource type and categories—making it easy to quickly discover useful and up-to-date materials.
🔍 Resources for Teaching CS
→ Computer Science II Lectures
Chris Bourke from the University of Nebraska–Lincoln is uploading his Computer Science II lectures to his YouTube channel. This is the first one. He previously uploaded these lectures during the Spring 2024 semester.
→ University of Virginia Teaching Hub
The University of Virginia Teaching Hub has recently published several new curated collections developed by educators from UVA and partner institutions. These resources cover a range of timely teaching topics, including strategies for supporting collaborative learning and group work, approaches to dialogue and deliberation across differences, concrete practices for fostering civil discourse—particularly during the first weeks of class—and innovative grading and rubric strategies for assessing student–AI collaboration. Together, these collections offer practical, research-informed guidance for instructors navigating contemporary classroom challenges.
→ Teaching Real-World ML Systems at Scale
This post is packed with concrete lessons for teaching real-world systems, data pipelines, and applied ML.
→ FossFLOW lets you draw isometric diagrams
If Excalidraw (mentioned below) feels a bit flat, this open-source editor allows you to create 3D graphics.
→ An open Algorithms Book
The Algorithm Codex is a free, open-source, literate-programming–style reference that explains undergraduate algorithms with intuition, proofs, and clear complexity reasoning—designed to complement textbooks.
→ 15.773 Hands-On Deep Learning
A graduate-level computer science course designed to equip CS faculty with up-to-date (well, up to two years ago) material on deep learning.
→ The Value of Agentic AI in the Research Lifecycle
These findings will serve as the foundation for a 2026 phase focused on the collaborative development of an open, technology-agnostic blueprint for high-priority agentic AI tools.
🦄 Quick bytes from the industry
→ Rui Fu’s Career Journey
Rui Fu (aka @raycfu) transitioned from a pre-med and biomedical research background into software engineering, advancing quickly through roles at Instagram and Meta. He now focuses on building in public while working on a stealth product. This conversation highlights four key ideas especially relevant for today’s computer science students and early-career software engineers:
Optionality over passion: Rui’s pivot into computer science wasn’t driven by a lifelong passion but by the desire for greater optionality. Coming from a pre-med and biomedical research background—and having struggled in AP Computer Science—his first real signal came when he landed a Nordstrom software engineering internship. There, he discovered that transferable building principles could scale across roles, eventually leading to Instagram and Meta.
Signaling over raw ability in recruiting: In Rui’s experience, landing a role wasn’t just about solving problems—it was about making reasoning visible in interviews. After a rejection from Microsoft, he realized the importance of articulating his thought process. He leaned heavily on referrals to escape the cold-apply pile and later used building in public to create visibility beyond résumés.
Judgment as the differentiator in the AI era: Rui views LeetCode as a screening proxy rather than a reflection of real-world software engineering. He emphasizes system design as the core on-the-job skill and recommends using tools like Excalidraw to sketch architectures and reason end-to-end. As AI-assisted interviews emerge (including at Meta), he argues that the most successful software engineers will be those who can question, audit, and reason about AI outputs rather than follow them blindly.
A nuanced take on the CS degree: Rui still encourages students to pursue a computer science degree, given that software increasingly underpins every field. However, he cautions that the degree alone is no longer enough—what matters is using it to get the principles down, especially as AI automates more routine coding work.
→ From India to Oracle
In this conversation with Amin Manazir, Shreyas Kulkarni explains why he left a comfortable life in India—less for “escape” and more for curiosity, opportunity, and a chance to experience a different culture—especially in a system he describes as intensely competitive and status-gated (where big tech recruiting often concentrates on IITs and top schools).
After studying CS at SRM IST Chennai and working at a fintech company, COVID delayed his plans, but he arrived in the U.S. and deliberately chose San José State University for one reason: ROI. By landing a TA role, he cut tuition from roughly $8.5k/semester to about $1.7k, earned a monthly stipend, and says he finished the master’s net positive (about $12k total tuition, plus earnings).
Being in Silicon Valley helped: SJSU’s practical curriculum and local recruiting pipeline led to multiple internship offers, including Oracle (with a signing stipend and strong hourly pay). He didn’t get a return offer from his intern team, but leveraged the internal recommendation to interview again and land a full-time Oracle role—later growing compensation substantially (he cites ~$300–$400k, much of it in stock).
Along the way, he shares a blunt playbook for breaking in—optimize for a parsable, keyword-smart resume, use referrals as a numbers game, practice LeetCode strategically, and “spoon-feed” interviewers clear evidence of ownership—while also acknowledging the tradeoffs: the U.S. can be high-upside but stressful, shaped by hustle culture, layoffs, and immigration uncertainty.
Notes I’m still thinking about:
Public vs. Private Universities (this section ties in really well with UH and my experience)
You don’t end up making 10× more money just because you paid 10× more for your degree.
In big tech, people from MIT, Stanford, public universities—all end up on the same team. The market doesn’t care how expensive your tuition was.
If your goal is employment, not prestige, public universities can be the smartest financial decision you make.
Brand names help you get interviews, not outcomes. What you do once you’re in matters far more.
AI-Assisted Interviews & Judgment
AI-assisted interviews aren’t testing if you can use AI—they’re testing if you know when not to trust it.
Companies aren’t looking for prompt engineers. They’re looking for software engineers who can reason when the model is wrong.
These interviews simulate real work: messy codebases, partial answers, and tools that don’t always help.
Interviewing even when you’re employed
Interviewing today is like training as an athlete—you don’t stop just because you won the last race.
The market is volatile enough that staying interview-ready is part of the job now.
Most people who get laid off didn’t do anything wrong. The only real control you have is preparation.
🌎 Computing Education Community
There are some really interesting tech talks happening here. If you live in Madrid or you’re passing through, don’t miss it!
Dwarkesh Patel is looking for remote scouts ($100/hour) to help find exceptional guests for his podcast.
The AI Council 2026 speaker application deadline has been extended to January 30.
CS faculty are invited to submit undergraduate course materials that thoughtfully integrate generative AI, with selected submissions featured at an invitation-only GenAI in CS Education Workshop in San Diego (March 16–17, 2026).
EduPar’26 invites papers, posters, and classroom-ready assignments on teaching parallel and distributed computing—including work integrating generative AI—co-located with IPDPS 2026 in New Orleans, with submissions due February 1, 2026.
This group invites faculty to join a project exploring how to intentionally develop professional dispositions in computing students through regular coursework, alongside knowledge and skills.
The Raspberry Pi Foundation is hosting a seminar on January 27, 2026, featuring Salomey Afua Addo on teaching neural networks and AI ethics in junior high schools in Ghana through unplugged activities and storytelling.
FIE 2026 invites abstracts and session proposals on engineering and computing education in the large language model era, with abstract submissions due January 26, 2026, for the conference in Paphos, Cyprus.
Anthropic and Teach for All are partnering to bring AI tools and teacher-led, community-driven training to educators across 63 countries, emphasizing global reach, local context, and teachers as co-creators.
🤔 Thought(s) For You to Ponder…
We’re living in an incredible time for content—there’s a nonstop stream of it every single day. At the same time, we’re completely flooded by it. For me, the real value lies in filtering, connecting, and making sense of it all within my own context. It also matters what kind of content we consume. We need to be critical, selective, and take an analytical approach. I try to do that every day, though it takes time and effort. I’m also fortunate in that I aim to make everything I consume useful in some way—whether personally or professionally. A beautiful reflection.
What does it mean for literature if machines can convincingly replicate a writer’s voice—or even improve upon it? This is an excellent piece from The New Yorker. It’s refreshing to see such careful attention to detail in an article like this!
+1TB. Over 20 million customers exposed (ID, bank info...). It took 2 hours and 30 minutes to steal the data. Self-taught young people. Fantastic radio report. The story behind the Endesa cyberattack.
Uncle Bob and John Ousterhout go head-to-head to solve—well, nothing. They just keep arguing, as usual. You can check out the follow-up discussion on GitHub or watch it on YouTube—whichever you prefer. BTW, I came across this podcast through the Stack Overflow newsletter, and it’s one of the few times I’ve felt the urge to binge the entire back catalog all at once.
This move by Accenture sends a very clear message: AI is no longer just an add-on to its portfolio — it’s the core of its next phase. Just look at who they appointed as CTO: the former CEO of Faculty, an AI-native company.
Paul Stamatiou’s 2025 Year in Review is a thoughtful, personal reflection on a year of major resets—leaving Limitless ahead of its Meta acquisition, taking time off, and joining Sesame—paired with sharp insights on craft, taste, and working with AI (especially his “year of Claude Code”). Concise, human, and refreshingly honest, it blends career reflection with the small details that actually shape how we live and build.
This isn’t the first time I’ve read that working with Claude Code is shifting people from being individual contributors (ICs) to agent orchestrators. It touches on the tension between efficiency and craftsmanship, pushing us toward a kind of meta-work that’s more like what managers do.
Are we suffering from “agentic psychosis”? I think it’s too early to say something like that — plus, addiction is a serious issue and not something to take lightly. But this article by Armin Ronacher really gives you something to think about: coding with AI agents feels productive, but does it actually degrade our judgment and the quality of the software? It also points out that sloppy PRs/issues put a burden on maintainers, not to mention the hidden costs and a culture of endless “loops” without proper quality control. So many questions. Honestly, I’m a bit confused. How about you? Share your thoughts in the comments.
The Bitchat app (created by Jack Dorsey) is allowing Ugandan and Iranian citizens to bypass internet shutdowns imposed by their repressive governments.
According to Affleck, AI ‘has no taste,’ and it’s taste that defines art. I wasn’t too far off in this past issue. I believe that the more omnipresent AI-generated content becomes, the more people will come to value real, human-made creations. And maybe that’s the whole point: to gain productivity and automate what we can, so we can focus on bringing the human perspective — the part that’s rooted in life and lived experience.
Written instructions remain, reduce cognitive load, and allow for repeated reference as needed.
📌 Research Corner
The spring semester started this past Tuesday! I’m excited to dive into the two courses I’m taking on campus—Research Methods and Advanced Numerical Analysis.
At some point this year, when my workload lightens up, I’d like to build a side project similar to Booklist—a curated collection of the most recommended reads online, but focused on tech books. Until then, you can check out the latest updates on the list in my Google Docs.
In order to improve my English skills (my goal is to speak like a native), I subscribed this week to this app built by Xiaofeng. It provides curated news articles tailored to my English level. Each article includes a cover image, vocabulary explanations, and interactive quizzes to help learners improve their English reading skills. I’m really enjoying the experience so far—let’s see if I stick with it!
Playing with Gemini 3 and Nano Banana to refine my TikZ graphics for the EDM paper this week—it’s an insane combo. For Cursor users, here are some tips on how to use the tool, shared by the CEO.
Teaching gets better when it’s done collaboratively, and watching others teach is one of the most powerful forms of professional development, as Mark Guzdial writes. I’m excited to start TA-ing Software Design next week under the supervision of my advisor, Amin! I can’t think of a more relevant course for the times we’re living in—where judgment, taste, responsibility, and domain knowledge remain irreplaceable.
🪁 Leisure Line
After having dinner there last Saturday, Vieng Thai is now my favorite Thai food I’ve had in Houston. Highly recommended—no affiliate link. We played Exploding Kittens afterward at home. Big fan.
📖📺🍿 Currently Reading, Watching, Listening
ORION is a short film about Orion Miller, a surfer in Montréal who braves frigid winter temperatures to ride the endless waves of the St. Lawrence River. It’s beautifully shot, and one surreal moment shows Orion floating on a large chunk of ice to get into position for the perfect wave.
Last Sunday, January 18, marked the beginning of the Week of Prayer for Christian Unity, which concludes this Sunday, January 25, with the Feast of the Conversion of Saint Paul. A divided Church cannot fully and authentically bear witness to the mission of Jesus. The lack of unity weakens everything. Unity isn’t the same as uniformity, but Jesus repeatedly emphasized how essential it is for us to be united in what truly matters. Take this example: Jesus chose twelve apostles who were very different from one another, yet He called them to unity — to row in the same direction when it came to what was important. It’s well worth remembering something Pope Benedict XVI once said: unity requires conversion. Unity doesn’t mean asking others to make up for what they lack — it means I need to contribute what I myself am missing: more humility, a greater willingness to listen, less of a need to impose my views or have the final word… so many things that require humility and the inner conversion Pope Benedict spoke of, so that true unity can shine through.
A necessary story. What a powerful episode. The risk involved in telling the truth about the Colombian armed conflict. Go here for the English version!
Seeking Beauty is a new documentary series now streaming exclusively on EWTN+. Hosted by actor and filmmaker David Henrie. Episode 1 focuses on Vatican City, and Episode 2 journeys through Rome, highlighting masterpieces, local voices, and how beauty points toward God.
🌐 Cool things from around the internet
🔗 Chromatic — A colorful daily puzzle game, created by Kevin Feyder.
🔗 flagstories.co — A striking visual breakdown of flag designs, created by ferdio.
🔗 Chosic — A free music discovery and playlist tool hub that helps you find songs, analyze your Spotify stats, generate names, and explore genres in fun and creative ways.
🔗 maptoposter — Transform your favorite cities into beautiful, minimalist designs.
Quick Links 🔗
🎧 Listen to Computing Education Things Podcast
💌 Let’s talk: I’m on LinkedIn or Instagram
If you’re enjoying Computing Education Things, please like, comment, or share this post! You can also support this work through Buy Me A Coffee. And if you’re finding value in the newsletter, consider forwarding it to a friend, colleague, or family member to help the community keep growing. Thanks for reading!
Didn't expect this take on the subject, but it makes so much sense, thank you for articulating so clearly that the 'magic was never in the code itself' as an implementation detail. Your point about the value shifting to clarity of thought and product insight, with engineers at the center of this transition, is a crucial perspective for anyone navigating the AI wave, especially for those of us teaching computer science.