I'm getting better (pre digested?) results for technical questions than google ever could. Last night it helped me with my redis replication info question, giving me the right commands to run.
Which is fine until it very confidently tells you things that are untrue, but look true (because it's good at presenting things that look true). It hallucinates entire libraries and meanings for command line parameters.
No, it's still fine because you can run commands and test them for yourself or read the docs. Just yesterday I asked chatGPT a question, then asked for a reference. It gave me a relevant paper to look at, and I discovered that it mis-quoted the paper.
It's fine as long as you don't take it at face value, just like interacting with a human :P