Although you’re being downvoted, I do agree with your perspective. ChatGPT, despite its limitations, has been a godsend for me.

I translate tons of texts to multiple languages, and ChatGPT not only provides pretty accurate translations as a starting point, but also allows me to see the pros and cons of specific translation choices. I can even notice recurrent ChatGPT mistakes and work my way around them programmatically.

I also heavily rely on it for making one-off scripts and programs to automate my workflow. Things that would require me to RTFM and waste time debugging syntax issues are no longer a problem. It does make mistakes sometimes, but it is often easy to fix them, too.

LLMs can’t be expected to be accurate, create original content or even solve hard programming tasks. But they really shine in certain workflows and that cannot be dismissed.