For work on operating systems, networking, storage, virtualization, or embedded systems in languages like C, C++, and Rust, where are AI tools useful and how much coding still needs to be done manually? Are they changing job prospects too?
How much is AI changing low-level systems programming work?
Answers
One useful role is a second pass over code for review and possible bugs, without relying on it to write production code. A suggestion can look plausible in isolation while missing the intent or constraints of the larger system.
That matters in low-level work, where the right implementation may be very specific to the problem rather than a familiar pattern.
Reply to this answer
Legacy integration is a different use case. AI can help rewrite code or replace library modules when an older C system has to connect to newer cloud services. That still leaves plenty of work in understanding the interfaces and how the systems fit together.
That kind of work may put more value on integration knowledge than on typing out the code itself.
Reply to this answer
For bare-metal and RTOS work, the code can look decent but the design may be more complicated than the task needs. If simplifying and checking it takes longer than writing the small routine yourself, there's not much of a win.
Reply to this answer
Job impact is harder to call. AI can take on enough coding work that fewer engineers might be needed for some tasks, but that is not the same as showing it can replace a systems engineer. Reviewing and correcting generated code still takes judgment.
There are different views here, from treating it mainly as a review tool to expecting a much bigger effect on staffing.
Reply to this answer
Add your answer
Answers are reviewed before publication.
_n48~zavqueno.7:
That distinction helps. A clean-looking implementation isn't necessarily a good fit if it adds unnecessary layers.