This matches my experience closely. LLMs are great at turning 10 minute tasks into 1 minute tasks. They're horrible at funding deep truth or displaying awareness of any kind.

But put some documentation into a RAG and it saves me looking things up.

Using some old tricks that used to work with gpt but don't anymore I was able to circumvent pretty much all censoring

https://i.imgur.com/NFFJxbO.png

So I'm finding it less censored than GPT, but I suspect this will be patched quickly.