These LLM code generation graphs have been making the rounds: https://lnkd.in/gGtdPt7f. I've used Claude Code the most with Elixir, Ruby, and C#. They all score relatively high in pass percentage (higher than Python, Go or JS, for example) in the charts. I haven't dug in to understand these results and what they represent well, but it has me wondering if this is one of the reasons I am having such a positive experience with LLM tools and yet I regularly talk to people who continue to have poor experiences with them overall. Just an off the cuff Friday thought.