I feel like LLMs are the death-nail for using anything but the top 5 most popular programming languages. It was already nigh-on impossible to make the business case for less-popular tools, simply from a hiring and training perspective, but with LLMs being trained on the most popular languages and tools, it doesn't matter how much better less common languages might be, they're not going to be able to overcome the productivity gain of throwing LLMs at the problem.