I used it extensively to debug a very thorny MySQL version upgrade. In my experience, it knows a lot about MySQL, and can reason about weird behavior very very effectively. A typical use is asking it something like, "Prior to my MySQL upgrade, my DB integration tests all passed, but now, they're non-deterministically failing with the following error. What could be causing this?" It then proposes hypotheses, which I either accept or push back on with new information. Surprisingly, this process led to it actually debugging a great number of my problems.

I use them every day for generating basic framework of whatever task I am doing, mostly coding. "Write an HTTP request that does this for this xyz test url that takes x y or z as input, create a table with the outputs, sample json attached." etc.

It saves me hours working on mundane shit every week.

90% of academics I know use chatGPT to help write grant applications or articles. Not for anything related to their actual domain, but more for improving language and clarity.

Proofreading, commenting, exploration, investigation, inspiration ...

Just because you are an expert does not mean you tackle every mundane but needed part of the work bare fisted or you memorized every edge case.

Programming is the one I see that comes up the most often. Unsurprisingly the experts are the best positioned to use it since they can identify hallucinations.

I frequently use it for mundane things. For instance, I had a set of go structs I wanted converted to rust. I told it to do the conversion using serde, and it spat out perfectly semantic rust. Doing this manually would have taken a hour, given the volume. This quite frequently works excellent for me... Complex, one off parsing problems

Writing code. Github Copilot is a big time saver.