This has not been my experience at all with gpt sol on ultra/max. Some concurrent code is just really really hard to write and even in those cases LLMs can help

This has not been my experience at all with gpt sol on ultra/max. Some concurrent code is just really really hard to write and even in those cases LLMs can help, the barrier for extensive thorough testing and writing your own custom simulator, liveness and correctness tests is simply so much lower now.