> While I certainly don't work at Cisco, I've come to the conclusion that I need to re-examine my priors when it comes to code rewrites and general maintainability.

One of the reasons to avoid rewriting has long been that the time spent rewriting code that works is time not spent writing code to meet new business needs. In other words, it's usually a huge opportunity cost, rarely balanced out by sufficient gains to be net positive.

I do think AI does challenge this prior, as the time needed to do a rewrite shrinks dramatically. It's not zero, of course, but it now becomes worth asking if it's worth doing, rather than defaulting to avoiding rewrites in nearly all cases.