@boilingsteam@mstdn.jp As a veteran developer, I will admit that I used code snippets, copied algorithms out of books, accepted auto-complete, and used canned routines. NPM and nodeJS have their attractions, but security isn’t one of them. I can see how AI search (not gen) could be part of a better faster secure development process. As a maintenance programmer, or supervising other developers, I don’t want the hassle of having to decode code that didn’t come from my juniors, and generalizing my testing for cases that ought not be programmed in the first place but could now be there. GenAI is like maintaining code made by the averaging of people you don’t know, not knowing how they think, assuming without confidence levels of competence you’ve grown to expect as professional. How could that be good?