The Computer Science Domain in the AI Revolution

Available on-demand

Witness the fundamental evolution of computer science through the eyes of Computer Science and Engineering assistant teaching professor William Theisen ’18, ’22 M.S., ’24 Ph.D. Learn why you no longer need to memorize the “punctuation” of computer languages, as the focus shifts toward designing the big picture of how programs are built. Because AI can now write lines of code almost instantly, your most valuable skill is knowing how to guide the machine and evaluating its results to catch “heinous” errors that a computer might miss. Understand why human oversight remains the essential final step in a world where technology can generate endless data but still lacks the wisdom to know where to “hit the hammer” to make things work correctly.

The New AI is sponsored on ThinkND by the Technology and Digital Studies Program in the College of Arts & Letters. This program collaborates with the Computer Science and Engineering Department and other departments around the University to offer the Bachelor of Arts in Computer Science, the Minor in Data Science, and the Idzik Computing & Digital Technologies Minor.

The Great Pivot: Pre and Post November 2022 The release of ChatGPT in late 2022 marked a “vibe shift” for the academic community. Dr. Bill Theisen, a professor and “triple Domer,” admits to initial skepticism that quickly turned to awe. While early models were niche curiosities, the subsequent arrival of specialized coding models fundamentally altered the landscape. For students like Matthew Chou, who entered Notre Dame just as these tools went mainstream, AI shifted from a tool for writing poems to an “agentic” force capable of handling entire programming assignments.

The Failure of Automation: A Case Study in Real-Time However, the promise of total automation often hits a wall of practical reality. Dr. Theisen shares a telling anecdote of academic adaptation: he recently spent eight hours attempting to prompt an AI to write a computer science exam. The result was “terrible”—a collection of flawed questions that a human expert could never accept. Ultimately, Theisen spent an additional hour rewriting the entire exam himself. This failure illustrates the limits of current models; while they can generate “cheap code,” they often struggle with the nuance and pedagogical intent required for high-level academic work.

The Evolution of Curriculum: Architecture Over Language Because “code is now cheap,” the Notre Dame curriculum is moving away from language-specific rote learning toward foundational logic and architecture. Theisen and Chou highlight that memorizing Python syntax is no longer a sustainable “moat.” Instead, the focus has shifted to an “abstraction level above language.” Students must understand the fundamental building blocks—functions, loops, and variables—so they can intelligently direct the AI and parse its errors. To maintain rigor, the department has pivoted toward “comprehension from yourself,” utilizing pen-and-paper exams and a shift toward oral examinations to ensure students truly grasp the material without a machine intermediary.

The Sociotechnical Challenge: The Social Deficit One of the more profound observations is the “social deficit” in modern CS education. Dr. Theisen notes that education has become “less social.” Office hours, once crowded hubs of collaborative whiteboard sessions, have been replaced by private interactions with AI “solutions manuals.” This erodes the communal learning experience; students often report having “no idea” what their project partners did the night before. This shift places a higher burden of self-discipline on the student to avoid the “brain drain” of offloading all cognitive labor to the machine.

The “So What?” Layer: Bifurcation and Design As we look forward, the era of universal access may be fleeting. Dr. Theisen warns of a “bifurcation of access.” We are currently in a “Golden Era” subsidized by venture capital, but physical constraints—specifically the massive demands data centers place on the power grid—may soon make top-tier models cost-prohibitive for the average user. In this future, the professional software engineer evolves into an architect of “formal verification,” focused on setting constraints and building testing frameworks for a trillion-line codebase.

How is the role of the software engineer evolving in a world of automated code? The role is shifting from a “writer” of syntax to an “architect” and “evaluator.” As generation becomes a commodity, strategic value lies in “formal verification”—setting the constraints and building the testing frameworks to ensure an AI’s output actually functions within a complex system.
What does “code is cheap” mean for the future of computer science education? It signifies that technical syntax is no longer a competitive advantage. Notre Dame is responding by emphasizing “comprehension from yourself” through oral exams and pen-and-paper logic problems, ensuring students understand the fundamental “why” even if they use AI for the final implementation.
How can students maintain learning integrity when an AI “solutions manual” is always available? The “social deficit” requires a higher level of self-discipline. Just as a solutions manual fifty years ago didn’t equate to learning, “watching” an AI solve a problem today is not a substitute for the cognitive struggle required to develop true engineering intuition.
Why is the “Golden Era” of AI access potentially temporary? Current access is subsidized by venture capital and faces looming physical limits. The immense pressure of data centers on the national power grid suggests a future “bifurcation” where state-of-the-art models may become cost-prohibitive for the general public.
Why is “logic and architecture” now more important than linguistic syntax? Effective direction of AI agents requires a deep understanding of fundamental building blocks at an abstraction level above language. Without mastery of logic (functions, loops, and variables), a user cannot parse error messages or intelligently direct a model to fix its own mistakes.

“The code, the hammer, [is] super cheap, but there’s a lot more that [goes into] it… It was like hammer one dollar, knowing where to hit the hammer like one thousand nine hundred and ninety-nine dollars. I very much think that still holds true.” — Bill Theisen
“Computer science has gotten less social, which may be hard to believe… It used to be if you were working on an assignment with someone, you were sat next to each other… Nowadays, I talk to students all the time, they’re like, ‘Oh, my partner did this last night. I have no idea what they did.'” — Bill Theisen
“You have to understand the limitations of it, as well as you have to understand the limitations of yourself… By having that holistic viewpoint, you’re not just handing off your thinking.” — Matthew Chou
“I think we’re actually in a bit of a golden era right now in terms of just like average Joe having access to close to state-of-the-art, and I think that changes.” — Bill Theisen
“If the models keep getting better and better and better, then like we’re all dead, right? … We all turn into paperclips.” — Bill Theisen