What’s a coding course for in the age of AI? Maybe not learning to write it
AI-generated summary reviewed by our newsroom.
- Professors predict coding may shift from manual syntax to architectural level thinking.
- NC State professors maintain no-AI policies while UNC’s professor predicts bans may end.
- Students are anxious but remain hopeful about employment prospects.
“Use of generative AI tools of any kind is not permitted in this course.” That’s a line from one UNC-Chapel Hill Introduction to Programming syllabus from the spring 2026 semester.
This kind of policy may soon seem like a relic from a different time in computer science education.
It may, in fact, be the last semester with any such categorical ban on the books at UNC-Chapel Hill, computer science professor Kris Jordan predicts.
In the near future, it’s possible that the majority of code may no longer be written by humans. That’s because of rapid advancements in artificial intelligence. Applications like Claude Code can spin up software based on prompts alone. It’s called “vibe coding.” By Google’s definition, vibe coding “marks the end of an era where software development required years of technical training, turning millions of non-coders into creators who can build and launch applications in seconds.”
You don’t need to know HTML to build a website. You don’t need to know Python to analyze data. You don’t need to know JavaScript to create an interactive app.
It’s like if you didn’t need to know how to read music to compose a symphony, or you didn’t need to master the elements of baking to make a layer cake.
Duke business school student Samantha Low says in a market research class she took, she was asked to do a lot of coding. The thing is, she doesn’t know to code. It didn’t matter. She was able to create tools to help her in her research anyway.
In the face of this transformative technology, computer science professors and students across North Carolina are considering existential questions about the future of their discipline. Once, Intro to Programming courses revolved around learning to write programming languages.
Now, it’s an open question: is there still a value in learning how to code?
The News & Observer spoke to professors and students at UNC-Chapel Hill and NC State University and reviewed more than 15 syllabi to find out.
The short answer? Yes, by all accounts, there is still value in learning the nuts and bolts of coding. But it may be a skill that gets used less and less.
Manual work of coding is disappearing
“Previously, we’ve had to teach coding as a sense of carpentry on your way towards learning architecture,” Jordan said. “[Now], we’ve got the ability to worry less about how to hammer nails, and more about the structure we’re putting together. We’re in a transition where we’re moving toward introducing students at a more architectural level of thinking.”
Students will be able to tackle bigger ideas and projects sooner, making for more exciting coursework, Jordan said.
NC State junior Sohini Das, who has interned with Deloitte and Google, says she’s watched a lot of the manual, syntactical work of coding disappear.
“A lot of the manual work behind the courses has now been replaced, but that’s been replaced with more critical and complex thinking,” Das said. Rather than endlessly correcting handwritten code, “we’re able to put more time into thinking about complicated systems or doing an extension on a project,” she said.
NC State computer science professor David Roberts is fascinated by the futuristic developments.
“The bottleneck is no longer how fast you can type, it’s how fast you can think,” said Roberts, quoting a Ted Talk by the creator of OpenClaw, an autonomous AI agent that can check your emails and book your flights.
But programmers and developers still need to know how to think, and what to think about, he emphasized.
Software engineers can be successful with AI because they know the fundamentals, they’ve grappled with the smaller problems, and now, they have a mental model of how different elements of computer science come together, argues Sarah Heckman, another NC State computer science professor. You need to know what you’re doing to successfully get AI to do what you want.
No-AI policies in intro courses?
Intro to Programming courses at NC State still have a no-AI policy. Heckman and Roberts are committed to that, unlike Jordan. The department is moving back to pen-and-paper exams for those early courses.
Jordan, Heckman, and Roberts all emphasized that even if the field becomes less and less about writing code, people will still have to know how to read it, test it, and work with it at a high level.
Instilling a philosophy around the use of AI in students is crucial for these professors.
“These tools represent one of the most powerful tools for learning in the history of mankind, but they’re also one of the most powerful tools for avoiding learning in the history of mankind,” Jordan said. “How you use them and engage with them makes a significant difference in your own personal growth and development. This is something that we’re hoping to train at a fundamental level in our students.”
Das says sometimes she takes a step back when using AI, and thinks: ‘Is this something worth learning myself?’ If the answer is yes, she’ll shut her AI down.
For Das and her fellow students, there is anxiety about the future of the field. But not hopelessness.
“What I want to do might be called something different, it might require different skills,” said NC State sophomore Ashika Kamjula. “It used to be that being great at coding could get your foot in the door. Now, because of AI, the expectations are different. Things like communication, system design, or thinking with a group of people: those things are just as important as being technically sound.”
“I’m not scared that my job is going to disappear,” she continued. “I’m just prepared that it might look different from what I thought a couple of years ago.”
Kamjula and Das both say that NC State is doing its best to prepare them for this brave new world.
