What happens when you start to trust AI-driven IntelliSense too much...

If you didn't catch the issue:

It's a stupid mistake, but it took me way too long to find it because I trusted Visual Studio so much.
The code didn't work and I focused on the cleanFileName instead.

@JS96 it flipped it, that son of a... or maybe a girl?

That's really dumb. I don't use AI intellisense though, only the autocomplete thingy (non-AI).

@Anchor I refuse. I might be conservative in this regard but the amount of dumb, unreadable or straight up incorrect code I have seen come out of this on both colleagues MRs or in general makes me refuse to rely on it.

Both that and I want to know what I'm doing, not just "looks good tab-tab-tab". I know using AI doesn't force you to mindlessly tab and you're still allowed to think about code but I like to avoid it all together.

@Anchor I agree it's useful sometimes (ex. create instant for loops, repeat the same line or function for N cases, etc.), but it should be disabled for suggesting arguments (leave only autocompletion), since it's clearly based on the type and not the meaning.
Too dangerous, there can be a lot of cases where the mistake isn't easy to detect and you find it only in a particular scenario. Something that helped you save 3 seconds, caused bugs you will have to detect and fix later losing minutes, or hours.

You must have forgotten to also accept its suggestion to overload the original function with the AI-correct order of parameters ?

well, this is one of those you can catch, I imagine with the influx of pasted directly from chatgpt code, the number of hidden bugs and heisenbugs will go through the roof

@jonathands heisenbugs

I like that.