Chatbots - Are They Secure Knowledge?

I think it's a huge overreaction personally because, yeah, this is not that big a deal. We could see chat to be making the source of mistakes pretty regularly, just making things up. Cora also announced a chatbot called PO, which is specialized for question answering. And now we have multiple companies like Flexi.i and maybe meta even going to be jumping into this soon. So it'll be interesting when you have half a dozen options for chatbots. What do you go with? And are they secure knowledge is I assume Cora is not training their own model like their own pre trained language model from scratch.