If I were a college student today, I wouldn’t pursue a CS degree. Maybe not even physics.
I’d probably go mechanical engineering with a heavy focus on robotics. Why? Because the second wave of the AI revolution isn’t going to live purely in the cloud or in prompt engineering loops.
It’s going to live in the physical world: energy-efficient hardware that can run inference at the edge without melting, actuators that respond with human-like dexterity, batteries and thermal management that make unconstrained autonomy actually viable, and embodied AI that interacts safely with messy, unpredictable reality (hospitals, battlefields, factories, lunar bases).
I’ve watched software engineering transform in real time, from a discipline where mastering multiple languages, deep systems knowledge (OOP, compilers, assembly, networks, OS internals), and careful architectural tradeoffs was the entry ticket… to something increasingly reducible to high-quality prompting, fine-tuning agents, and reviewing/steering agentic outputs.
None of that trajectory screams “you need four years of formal CS lectures to succeed.” My generation might be the last one that had to grind through writing a compiler from scratch or debugging race conditions in bare-metal code. In the next couple of years, fully autonomous agentic full-stack development (frontend to infra) feels not just plausible, but probable.
The current CS curriculum in many universities feels like teaching punch-card programming during the dot-com bubble. Valuable history, but not the sharpest tool for tomorrow’s leverage.
If you’re a student, don’t major in only computer science. Think three steps ahead. Pick a domain that gives you a competitive edge when software development becomes a commodity. Hardware that makes AI physically useful. Robotics that turns abstract agents into real-world impact. Interfaces between bits and atoms.
Medicine (surgical robots, prosthetics), healthcare (exoskeletons, assistive devices), defense (autonomous systems that actually survive chaos), aerospace (lightweight, extreme-environment robotics)… these are where taste, domain depth, and systems intuition still compound massively.
We’re not just watching the future unfold. We’re the ones who get to shape where intelligence meets the physical world.