After testing Claude and Grok and found they give around 6-7 wrong answers out of 10 questions.
ChatGPT update from this morning:
5 wrong answers out of 5 questions.
LLMs are just a waste of time.
LLMs are the biggest farce we have seen since the .com boom.
They are trained on internet slop garbage. Trash in trash out.
At best they are application specific and not broad overarching tools.