Of course, without fluency in the language that your AI agents are using... good luck trying to assess their fluency. ;-)
I get the point Hamilton is trying to make, but I still feel more comfortable using languages where I have deeper expertise, especially for "real work".
Your fluency in a programming language matters a lot less now. AI's matters more.
Pick:
- Mainstream lang (so there's enough training data docs)
- Good attributes (perf, tooling, types)
Then focus on the product and architecture and let the agents run wild.