I'm really at stump here, I want to use AI to make my life easier with work but how do I if I can't trust its answers 100% of the time? I mean I understand it needs human intervention but if I don't know the answer itself, it involves me stopping at the exact time I need to, searching for the correct answer elsewhere entirely and or just giving up on it, its quite exhausting.
The process for me is usually I go down a path with an LLM; I get excited at the prospect of using it as a tool. In the middle i spot an error. Question it, the LLM responds with a contradiction in the same conversation. I realise it means the last 5 minutes of dialogue was all incorrect, and I get frustrated that it wasted my time. I mean they're selling it as intelligent but where is the intelligence, to me its more akin to a search engine where instead of showing me a full article, it returns the excerpt of text instead. Useful.. maybe?
I know people use this in their work daily so I'm just curious how do they do it? Do they ignore the blaring errors like they aren't there? I know this sounds sarcastic but I really want to know because especially in my own projects, I find it hard to move on from not 'perfection' per say, but anything other than bad.
Thank you in advance and looking forward to other opinions also
ungreased06755 hours ago
What kind of work are you trying to do?