Every couple of months, someone on the internet proclaims that learning to code is a waste of time because artificial intelligence will soon write all software anyway. But when the man widely recognized as the “Godfather of AI” — Geoffrey Hinton — says that computer science degrees will continue to have value for a long time, they put down their phones and pay attention. He’s not in the business of giving comfort quotes.” If anything, he’s spent the past couple of years warning the world about what AI could go wrong more than what it could do right.
So when young people are told by Hinton to learn how to code, it’s different. He’s not romanticizing the past or defending traditional education. He’s being practical. He understands how these systems function at the most fundamental levels, because he helped build the platforms upon which they operate. And, in his view, AI is not rendering coding irrelevant — it’s making it more important.
You watch him talk about it, and you can understand the logic. AI tools can write code, however they don’t understand problems the way people do. They don’t have a strategic way of thinking about system design or system architecture. They don’t calculate tradeoffs or project long-term consequences. They receive prompts and return outputs, and the quality of those outputs is almost entirely determined by the human providing instructions. You have to know what you’re asking for, or you won’t even know when the A.I. gets it wrong. And it will get things wrong.
That’s why Hinton keeps advocating that students must continue to learn genuine computer science — not just how “to ask AI for help.” A generation that completely leans on automation without knowing how it all works, is a generation potentially sitting duck to subtle bugs and giant security holes and systems that look great until the second they totally don’t. It’s often said that knowing how to code is a superpower, just like being able to reproduce the few strokes it took Leonardo to paint the Mona Lisa. It’s really about thinking in systems and logic and cause-and-effect.
There is also the reality that AI has made coding more accessible, not extinct. With these tools, beginners can make things faster, try out ideas quickly and learn by doing. But the people who benefit most from AI coding tools are those who already know the craft. They know how to debug. They can often tell when there’s something that AI is not making sense to them. They know what works and what is just disaster posing as a solution.
Hinton sounds legitimately concerned about all the kids who believe this AI thing will just go and do it for them. In his eyes, that’s a trap. The future belongs to people who understand AI, not those who treat it like a magical distraction. And you can’t grok AI without getting programming, algorithms, data structures, mathematics and the logic that holds it all together.
He’s not alone in that view. Tech leaders throughout Silicon Valley have been echoing the sentiment: AI reshapes the workflow, but it doesn’t eliminate the need for technical know-how. If anything, it sets the standard. Engineers aren’t even going to spend their days writing boilerplate code. They’ll use them to design systems, debug AI-generated logic, craft prompts that actually work and build the infrastructure that automates everything else.
The short version: AI makes coding a job higher up the development chain, not one that goes away.
Hinton also realizes that invention doesn’t result from just employing existing tools — it comes from making new ones. We won’t see the next wave of AI breakthroughs emerge from prompt writers. They will come from people who have a sense of how models work, where they fail, how to train them, how to optimize them and also for whom it makes sense to take cutting-edge technology and push it beyond its current limits. And that’s not something you can fake. It’s not something an A.I. can give you. That’s computer science.
And when he says CS degrees “will be valuable for a long time,” what that also means is something more profound: The world will still need people who know how to understand the machine. Not how to use it but how to query it, fix it, question it and build on top of it.
AI might automate tasks. It might change workflows. It might reshape entire industries. But because humans, not just machines, will continue to be the architects of such systems, coding isn’t going anywhere — it’s a necessary skill if you want to help design the future rather than be its victim.
That, in part, is why Hinton’s advice is so searing. He’s saying to young people: don’t just sit on the sidelines. Learn the tools. Learn the logic. Learn the foundation. AI is not replacing you — unless you understand it.
