It’s not even about trust. There are not really AI’s s they have a problem with fixing one-line bug where solution is in official documentation. I can’t think about it to do something bigger for me. I sometimes use it for a hint or idea and research specific keywords/naming by myself and that’s all we can really do with so-called AI’s. As always the end of the humanity/world has been delayed …

I have just tried it and results were not surprising. There are 2 different levels with and without searching internet … I would say that the one without searching is on Chat GPT-3 level and the second one is on Chat GPT-4 level. However both answers were completely wrong. When read carefully and browsing search results they give a hint, but it’s at least as same amount of work as without AI, so they don’t help me at all.

Edit: I have changed one thing in my example 1 LOC code and the results with and without internet were terrible. They have caused more problems than help. This AI sometimes cannot tell a difference between language types.

I wonder if my 1 LOC test is better than Turing’s one as so far no AI have passed it at all.

Just using ChatGPT to help me once in a while, specially with regex stuff

I found MS copilot sometimes not giving accurate information.
It tried to get code from forums and try to code by itself, but sometimes the code is not working or having syntax errors.