Why Geoffrey Hinton Says a CS Degree Still Matters in the Age of AI

Universities across the U.S. are seeing a sharp shift: enrollment in traditional Computer Science (CS) programs is falling, even as AI hype surges.

According to our previous report on the decline in CS majors, about 62% of computing departments reported a downward trend in undergraduate enrollment this fall. Meanwhile, dozens of educational institutions like MIT now offer standalone AI- or ML-focused majors. Many students are switching, betting that AI credentials may give them a fast track into lucrative tech roles.

But before you close the door on CS in favor of “the next big thing,” consider this: the “godfather of AI,” Geoffrey Hinton, argues that a CS degree remains uniquely valuable. And you might be making a mistake by skipping it altogether—here’s why.

Hinton’s core point: a CS degree isn’t only about writing lines of code. It’s about building deep foundations in mathematics, statistics, systems thinking, and conceptual clarity, all of which no AI can replicate (yet).

In an interview with Business Insider, Hinton likens learning to code through a formal CS program to learning Latin. While it’s rarely used directly, it’s crucial for sharpening thought, logic, and understanding of structure.

Hinton isn’t alone. Other respected voices in tech are echoing a similar sentiment that in the age of generative AI, what matters isn’t how many scripts you can pump out. What truly matters now than ever before, is whether you understand how and why those scripts work; whether you can architect systems, reason about complexity, and solve hard problems.

A CS degree delivers more than syntax and APIs. It provides problem-solving frameworks, algorithmic thinking, rigorous logic, and system-level understanding. Regardless of the industry you find yourself in, those are the skills that matter when you’re not just using tools, but building and evaluating them, or working on infrastructure, security, or AI systems design.

As evidence that shallow AI-assisted coding can have downsides, consider the recent internal study by Anthropic. The 2025 report found that their engineers reported that roughly 27% of tasks are done with AI-assistance.

While coding with AI resulted in engineers’ productivity gains, many also warned of “skill erosion.” Engineers said they spend much more time debugging and cleaning up AI-generated code than writing original logic from scratch. It also had adverse effects on the talent pipeline, as junior engineers and their senior counterparts were less likely to collaborate, whether to work through errors together or simply ask questions regarding code quality and accuracy.

That appears to be the core danger of merely relying on AI to generate code. You risk losing the deeper understanding of why it works, or why it doesn’t.

Someone with a strong CS background, though, will be better equipped to spot issues, understand architectures, and build robust, maintainable systems. As Hinton suggests, that foundational training may become more, not less, important, especially as tools get more powerful and code becomes more opaque.

If you’re a student hesitating between “should I major in CS?” or “should I jump straight into an AI-focused program (or skip college altogether)?” the current trends can look discouraging.

The dip in CS enrollment may seem like a collective vote of no-confidence. But maybe it’s just fear or a short-term reaction to massive shifts—not a reason to abandon long-term thinking.

Hinton’s take argues for a different perspective. Instead of treating a CS degree not as a quick ticket to a coding job, think of it as a long-term investment in intellectual tools that stay relevant even as industries change. If your goal is more than writing a few scripts, from designing systems and understanding AI internals to tackling complex, real-world problems, then CS may be more valuable now than ever.

That said, given how fast the job market is shifting, flexibility matters. It might pay to complement a CS education with adaptable, interdisciplinary learning. It helps to consider enriching the traditional, formal CS route by dabbling in data science, AI, cybersecurity, or even non-traditional fields like computational biology, hardware, or AI policy.

And if you want to be job-ready when you graduate, combining formal education with practical exposure can help a lot. For example, Interview Query has a wide range of resources from interview prep guides to coding challenges and learning paths, all of which can give you hands-on skills that pair well with the broad foundation of CS theory.