Good answer, for sure. The one thing that is dangerous is that it's hard to know, when you're going off and learning "esoteric" skills, if those will be winning horses or dead ends. A lot of people would rather max out on enterprise Java (which has a well-studied, if commoditized, market) than take a risk on a specialty that might be "hot" today but dead tomorrow.
The AI winter is one of the worst things to have happened to software, and the decline of funding and even respectability of basic research has set technology itself back 25 years, and there's a scary lesson in it, which is that interesting, cutting-edge work can suddenly enter a funding drought.
The AI winter is one of the worst things to have happened to software, and the decline of funding and even respectability of basic research has set technology itself back 25 years, and there's a scary lesson in it, which is that interesting, cutting-edge work can suddenly enter a funding drought.