I gave a talk in December 2016 titled “Chatbots in 2017” which was well received but was hardly prescient. We are hearing now that AMZN is not happy with Alexa as a business even though I think many people like it.

So we’re left with the feeling that the chatbot UI is not successful and people are waking up to GPT-3 as not the path to Skynet but something that spits out BS while following the form of a certain literature.

I really want to see something better but we seem stuck. I have to admit I liked Clippy but that was because it had a well tuned search engine and good documentation behind it which is not usually the case for online help. It could be the theming and characterization turned people off to a very good help facility as it turned them on.