Once people try to increase quality instead of speed they will see how LLMs are powerful. Everything else is just sales pitch by Nvidia and friends.
Debugging and developing first fixes is also one of the spaces where current LLMs are the biggest force multipliers. Especially if you have reproduction cases the LLM can test on its own
But long-term it might look very different as more and more of the code becomes LLM written