Some of these are still alive in AI as well, although often under parallel names. For example, genetic programming (an old field of AI) has broadened to be more

Some of these are still alive in AI as well, although often under parallel names. For example, genetic programming (an old field of AI) has broadened to be more method-agnostic and doesn't really care if your automated programming method is "genetic" anymore or not, so has converged to some extent with program synthesis. As a result you can publish this kind of thing in both PLs and AI conferences nowadays, with a bit of translation. (The overlap of people who actually do publish in both types of venues is less than I might hope in an ideal world, but a handful of people do.)