Increasingly I’m realizing that in most cases there is a SIGNIFICANT difference between how useful AI is on greenfield projects vs how useful it is on brownfield projects. For the former: pretty good! For the brownfield, it’s often worse than useless.
Oh, I find almost the exact opposite.
On Greenfield projects there's simply too many options for it to pursue. It will take one approach in one place then switch to another.
On a brownfield project, you can give it some reference code and tell it about places to look for patterns and it will understand them.