Evaluating Developers in the Age of AI

First of all, a developer solves a certain amount of problems, and definitely not all of them are connected to the framework he or she uses.

I would argue that most of the problems a developer solves are domain- and architecture-related (that is, what architectural aspect we're working on, the way our modules communicate, etc.). Then, the second part of the problems is the language we use itself, because languages have their limitations, nuances, ecosystems, etc. And only a small chunk of things is related to problems created by the frameworks you use.

In the modern world, problems that are tech-stack related will go away. When I use AI to write code, I rarely encounter problems related to the framework, but I always need to think about code quality, the architectural decisions I make, and the domain rules that need to be preserved.

So basically, what I'm saying is that, for example, people who want to find a good developer don't need to pay attention to his main framework, but rather to his main language, his field of tech expertise (backend, frontend, security, distributed systems), and his domain-specific expertise.

Now, almost every developer will be able to solve problems he or she is aware of using AI. The main problem is that sometimes developers don't know about the existence of specific problems in the first place, so they omit them entirely. And that may result in dire consequences.