No need to hit the panic button as yet. Large Language Models (LLM's) are confident bullshit generators. You should only worry about them replacing you if you're working mostly on low-value code (i.e., simple problems). They sometimes fluke their way into higher value code, but I wouldn't touch it with a barge pole. Will they get better? Yes, for sure. But guess what ... this only means you'll stop wasting your time doing low-value code and focus on solving harder problems. Until ( if ever) we hit the singularity, you really don't need to worry about AI at the level of LLMs.