There are plenty of seats and dim lighting in a basement classroom on the Stanford campus. Standing at the front of a room full of computer science students, Mihail Eric, a Stanford alumnus who is now a lecturer, tells them something that hardly any professor in the nation would say aloud: he will teach them how to create software without ever writing a line of code. No one chuckles. Some of them appear genuinely relieved.
At one of the world’s most competitive universities, Eric’s course, The Modern Software Developer, has quietly emerged as one of the most popular courses. It presents itself as the first significant effort by a major university to integrate AI coding tools like Cursor and Claude into the curriculum rather than merely tolerate them. Stanford is taking a different approach at a time when the majority of academic institutions are still debating what constitutes cheating and drafting AI use policies. It is inclined inward.
It’s difficult to ignore the timing. Even at a university like Stanford, this is an exceptionally stressful time of year for computer science majors. One of the students in Eric’s class, Brent Ju, is graduating this spring without a job offer. The market, according to him, is “a little tough.” It’s a statement that most likely downplays the circumstances. AI now writes about 30% of Microsoft’s code, according to CEO Satya Nadella. Earlier this year, Dario Amodei, CEO of Anthropic, went so far as to say that within a year, AI might be in charge of “essentially all” of his company’s code. It’s hard to accept for someone who spent four years and an incredible sum of money learning to do just that.
When Eric graduated from Stanford in 2016, a computer science degree from a prestigious university seemed to be a surefire way to land a lucrative tech job. The conventional wisdom was simple: get the degree, get the job, stay comfortable. The FAANG companies were hiring aggressively, and salaries were rising. That computation has changed dramatically. Overstaffing during the COVID-19 pandemic was followed by a hiring freeze, a wave of layoffs forced seasoned engineers back into the workforce, and AI began performing jobs that previously required junior hires. The math stopped functioning.
In a sense, Eric is answering that math directly. Instead of treating AI tools as academic dishonesty or as nonexistent, his course views them as an integral part of the workflow in a professional engineering setting. Boris Cherney, the man behind Claude Code, and Silas Alberti, the head of research at Cognition, were among the guest lecturers. Alberti gave a talk titled “The Opinionated Guide to AI Coding in 2025.” Following his lecture, students gathered around Alberti in a manner similar to that of a visiting athlete. His message was clear: “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.”
Walking through Silicon Valley at the moment gives the impression that the industry has already decided. Hiring managers at large corporations are assessing applicants based on prompt engineering, their capacity to examine and fix code produced by AI, and their comprehension of the shortcomings of AI. Practically speaking, a student who graduates without ever using these tools is behind. All Eric’s class is doing is recognizing what the industry already knows.
Whether this model will spread rapidly or continue to be an anomaly is still unknown. Due to worries about academic integrity and mounting evidence that AI literacy is becoming a standard requirement for professionals, the majority of universities are proceeding cautiously. Without prohibitions, the more difficult question becomes what students actually need to learn on their own, and there is currently no clear answer to that question, according to Stanford education researchers.
Standing outside that Palo Alto basement classroom, it’s obvious that something is changing. The students leaving after Eric’s lectures don’t appear to have been given a free pass. They appear to be individuals attempting to understand what it means to be useful in a field that is redefining usefulness more quickly than most institutions can keep up.
