Architecture decisions, requirements analysis, trade-offs in technology selection, and cross-system debugging—these high-level cognitive activities cannot yet be replaced by AI. Focusing on these areas is actually a smarter skill investment.
They can be easily replaced. I suspect we are just engaging in a form of self deception by thinking these choices we make are more than arbitrary decisions driven by nothing but our taste...
Often times I find agreeing to the stack the model suggests is the right option. It’s probably most trained on that.
I suppose one would have to work in a very specific novel and niche area to go against the grain and chose stacks that are not chosen by majority.
Do you mean any other skills can be actually replaced?