In a dimly lit Stanford University basement classroom packed with anxious computer science students, lecturer Mihail Eric tells the class he’s going to teach them how to code without writing a single line of code.
Eric’s class, The Modern Software Developer, has quickly become one of the hottest Stanford CS courses this semester, which bills itself as the first attempt at a major university to embrace coding tools like Cursor and Claude.
It is an unsettling time to be a computer science major, even at a school as prestigious as Stanford, knowing you will be graduating into a world where AI is getting better at programming by the day.
“It can be scary because you think your job security is being compromised, and you might get replaced,” said Brent Ju, one of the class’s dozens of students. Ju is graduating this spring and so far has no job offers. “The market is a little tough. I am still interviewing.”
“If you can go through this entire class without writing a single line of code, more power to you,” said Eric, a Stanford alum who purposefully designed the course to be an antidote to the majority of classes that still ban the use of AI.
A who’s who of AI coding luminaries has stopped by the bucolic Palo Alto campus to guest lecture, including Boris Cherney, creator of Claude Code, and Gaspar Garcia, head of AI research at Vercel. Martin Casado, a general partner at Andreessen Horowitz, will address the final class next week.
On a recent morning inside the classroom, Silas Alberti, head of research at Cognition, delivered a lecture called “The Opinionated Guide to AI Coding in 2025.”
Ben Bergman/BI
“I think what you learn in school has always been a little bit behind, so I’m glad that this course exists to teach the newest stuff,” Alberti said after his lecture, surrounded by students lined up to greet him like a celebrity. “If you learn with yesterday’s methods, you are not going to be super competitive, but if you really lean into the tools, you can be a super engineer.”
Excitement and fear
The mood of the students in the class reflects the current zeitgeist of Silicon Valley, with excitement about what many consider one of the most significant technological advancements of our lifetime. But there is plenty of fear about AI rendering an expensive Stanford degree obsolete.
When Eric graduated in 2016, getting a Stanford CS degree was the golden ticket.
“People thought ‘I’m going to go to an elite university, and then I’m just going to be set for life and have a cushy six-figure job for as long as I want at a FAANG company,'” he said.
The number of CS students surged as tech companies embarked on a massive hiring spree.
“Meanwhile, a lot of companies that hired a lot during COVID saw that they overhired,” Eric said. “Now you have a surplus of young talent and also a surplus of newly laid-off, quite experienced talent.”
Making matters worse, AI is already proficient in coding and continues to improve rapidly. Microsoft CEO Satya Nadella has said up to 30 percent of the company’s code is being written by AI, while Anthropic’s CEO Dario Amodei predicted in March that AI had the potential to write “essentially all” of the company’s code within a year.
Ju, who says his dream job would be to work at Anthropic, says he is trying to stay positive.
“It’s exciting because if the tools aren’t going to replace you, but act as an assistant, it can really supercharge your productivity and make you a more effective developer,” Ju said. “I’m more of an optimist who leans toward that direction.”
Zach Lloyd, founder and CEO of Warp, a developer tool for agentic workflows, delivered a guest lecture last month and maintains he is still very interested in hiring CS students.
“The idea that people from a place like Stanford with a CS education won’t be able to get jobs as engineers is a little overblown,” he said, adding that knowing the fundamentals of programming is still vital to effectively using Warp or Claude. “These tools are accelerators but not replacements yet, and the actual people who will be best at wielding them are those who have a solid foundation.”
Eric plans to teach the course again next year, though he says AI is advancing so fast that the class will likely look very different.
“People were asking me if I was concerned that by week seven, things are going to be obsolete that I talked about in week one?” he said. “Yes, it is a concern. So far it hasn’t happened yet.”
