They did everything to learn to code, then tech moved on to AI
After a decade of promises about high-paying tech jobs, recent graduates often cannot get interviews while schools race from coding contests to AI contests.
The Vergecast · Schools are catching onto tech’s playbookOne of the recent computer science graduates interviewed by Natasha Singer, the New York Times reporter who covers education and technology and the author of the book “Coding Kids,” was in first grade when every adult he met gave him the same formula for the future. Just learn to code, they said, and you will have an amazing job in tech.
He listened. He learned HTML, the basic language used to build websites, and was making his own sites in elementary school. In middle and high school, he made apps. In college, he majored in computer science and served as a teaching assistant for the subject. He did not do it only for a paycheck. Singer said he loved “the logic and the puzzle of programming.”
Singer spent months last year talking to recent computer science graduates who could not land work and often could not even get an interview. At the same time, tech giants were investing hundreds of billions of dollars to build data centers for AI while laying off workers not involved in AI.
Then came the new advice. The graduates were told it is easy for them to learn AI, to use AI tools to write software, or to pivot and become what the industry calls forward-deployed engineers, consultants who help other companies put more AI in their businesses. To some students, that felt like a betrayal. Some of the students said, “I feel like the tech industry gaslighted me, and I feel devastated that I devoted all these years to this thing that, you know, nobody cares about anymore.”
Schools are feeling the same whiplash. The nonprofit Code.org, which helped bring coding lessons into public schools, is not abandoning computer science but is expanding AI’s role in how it teaches the subject. Singer says schools moved from code-a-thons, contests built around writing code, to prompt-a-thons, contests built around giving instructions to chatbots. Now, because people can simply talk to AI, the push is for every kid to learn vibe coding, describing software in plain words while AI writes it.
Singer asks what we owe students who were promised “high-paying, high-powered tech jobs” for a decade. Singer says the Computer Science Teachers Association, which speaks for 10,000 teachers, “has decided that computer science should center humans, not machines.” Her question is how to teach kids durable skills that work across any technology, whether it is apps, social media, AI, or quantum computing, the super powerful computers that may come next.
Questions people ask
What is vibe coding?
Vibe coding means describing the software you want in plain language and letting an AI system write the code. It works well for simple apps and prototypes, but professional programmers still need to check the results for errors, security problems, and legal issues.
Are entry-level coding jobs actually disappearing?
Hiring for entry-level software roles has been unusually weak since 2023, and many companies say AI tools let smaller teams do more routine coding. There are still jobs for new graduates, especially in AI-related work and cybersecurity, but competition is much tougher than students were led to expect.
What do teachers mean by centering humans, not machines?
It means teaching students to understand how computers work, how data is used, and how technology affects society, rather than just training them to use a company's current product. The goal is judgment about when to trust technology and what its tradeoffs are, a skill that carries over when products change.