The larger point is that by keeping non-strategic part of your tech stack boring, you can focus on core innovative technology. In this guy's case it's LLM.

Which is same point that the original post "choose boring technology" makes. I don't see the contradiction.