LLMs have massively increased the value of code that can be: - locally reasoned about (pure functions that compose) - static type checking (so that nonsense i

LLMs have massively increased the value of code that can be: - locally reasoned about (pure functions that compose) - static type checking (so that nonsense is discovered at compile time) Weirdly this makes Python a worse choice and Haskell a better one. Clojure benefits a little bit. I don’t think people are choosing language stacks based on what is best for the problem. The main factor is “what do I already know?”