Does CS as a Degree Still Matter in the Age of AI?
My reflections and perspectives on being a CS student at the University of London while building an AI app studio
There’s a particular kind of cognitive dissonance that comes with being a Computer Science student in 2025.
On one hand, you’re sitting through lectures on binary trees and asymptotic complexity, carefully annotating your notes like you’re preparing for an exam that, frankly, feels increasingly disconnected from the world outside the classroom window. On the other hand, you’re watching seventeen-year-olds ship full-stack applications with nothing but a Claude subscription and sheer audacity — no formal education required, no prerequisites, no gatekeeping.
So the question becomes almost unavoidable: does the degree still matter?
I’ve been living this question in a very literal sense. I’m currently a CS student at the University of London (Goldsmiths), and simultaneously, I’m the CEO and co-founder of Mindify AI — an AI-native advertising infrastructure company I built alongside two brilliant co-founders I met through the NTU AI community in Taiwan. I spend my mornings reading about graph traversal algorithms and my afternoons trying to figure out how to close our seed round. I’ve had papers cited by researchers I deeply respect, and I’ve also sat across the table from investors who couldn’t care less about my h-index.
That tension has taught me something I didn’t expect to learn.
The fundamentals aren’t about coding. They never were.
When people ask whether CS still matters, they’re usually asking the wrong question. They’re thinking about syntax, about languages, about whether a degree teaches you React or Rust or whatever framework Twitter is collectively excited about this week. It doesn’t, and it shouldn’t.
What CS actually teaches — when it’s taught well — is how to think about problems structurally. Data structures aren’t just academic trivia. They’re a framework for understanding how information wants to move, where it gets stuck, and why some solutions scale while others collapse under their own weight. Algorithms are really just a formal vocabulary for reasoning about efficiency and trade-offs — two things that matter enormously when you’re building a real product with real constraints and a burn rate that keeps you up at night.
I noticed this acutely when our CTO Cheng-Peng and I were working through the architecture for Cortex, our AI chatbot with integrated ad delivery. The decisions that felt most natural to me — the ones where I could argue confidently about why one approach would perform better at scale — were grounded in things I’d learned in lectures I’d half-resented attending. Computational thinking turns out to be one of those invisible muscles. You don’t notice it until you need it, and then you’re very glad it’s there.
So yes, CS still matters. Not as a credential, but as a mode of reasoning. The fundamentals are evergreen precisely because they’re not about the technology — they’re about the logic underneath it.
But here’s the uncomfortable truth: CS alone will get you nowhere interesting.
The most valuable thing I’ve built in the last year isn’t a feature or a product — it’s a mental model that sits at the intersection of technology, consumer behavior, and market dynamics. That model doesn’t come from any one discipline. It comes from reading obsessively across fields, from talking to advertisers who’ve never touched a line of code, from studying how attention actually works in psychology, from understanding regulatory landscapes and investor incentive structures.
We’re in the age of multidisciplinary everything, and the people who will define the next decade of technology aren’t going to be the ones with the deepest specialization in any single domain. They’re going to be the ones who can hold multiple frameworks simultaneously — who can read a machine learning paper in the morning and a behavioral economics study in the afternoon and synthesize something new from the friction between them.
My co-founder Stella, our COO, approaches problems from angles I genuinely wouldn’t arrive at on my own. That’s not a weakness in our team’s coherence — it’s the entire source of our strategic differentiation. The CS background gives us a shared technical foundation, but it’s the divergence in how we each think beyond that foundation that actually creates something worth building.
If you’re in a CS program right now and you’re only reading CS, I’d gently suggest you’re leaving the most interesting part of your education on the table.
And then there’s the thing nobody in academia talks about enough: agency.
This is the part that I’ve come to believe matters more than almost anything else.
In the age of AI tools, the raw cost of building has collapsed dramatically. You can spin up infrastructure, generate boilerplate, prototype a feature, draft a pitch deck — all of it happens faster and cheaper than it did three years ago. That sounds like pure democratization, and in many ways it is.
But here’s what that shift actually does: it removes every excuse for inaction. It strips away the friction that used to justify hesitation. When the tools are this good and this accessible, the only remaining differentiator is whether you actually do the thing — and how relentlessly you pursue it.
Agency, I’ve come to believe, is the ultimate multiplier. It’s the variable that amplifies everything else in the equation. A team with modest technical skills and extraordinary agency will consistently outperform a technically brilliant team that moves cautiously, waits for permission, or second-guesses itself into inertia. I’ve seen it. I’ve felt it from both sides.
Building Mindify AI has been a crash course in this reality. There were months where the uncertainty was genuinely vertiginous — zero revenue, competitive pressure from companies with fifty times our funding, questions we couldn’t yet answer. What kept us moving wasn’t certainty. It was the cultivated habit of doing the next concrete thing anyway, even when the full picture wasn’t clear. That habit — that instinct toward motion — is something no curriculum teaches you directly. But the best CS programs, I think, create environments where you can practice it, if you’re paying attention.
So does CS still matter?
Yes — for the thinking patterns it builds. No — if you treat it as sufficient on its own. And ultimately, it matters far less than what you decide to do with it, and how urgently you decide to do it.
The degree is a starting line, not a destination. The most important question you can ask yourself isn’t what did I study? but what did I actually build, and how hard did I push?
The tools have never been more powerful. The question is whether you’ll pick them up.