It really sounds like you have not personally used ChatGPT+Bing, right? I have been using it as my search engine for a few days and when in chat mode, it shows

It really sounds like you have not personally used ChatGPT+Bing, right? I have been using it as my search engine for a few days and when in chat mode, it shows you the web search queries it uses to gather context text to feed to ChatGPT along with your question. It then provides reference links. Just one case from yesterday: we had been considering leaving our parrot by himself for two nights. A few times a year we leave him for one night and he seems to do OK. I asked ChatGPT+Bing about this and I got a very good response on why leaving him by himself would be a bad idea. I checked the references it used, and then hired someone to bird sit. I have also asked about business ideas, and other topics that I would not have thought to ask a search engine previously. At the heart of it, Squad type deep learning models have for years done a good job at answering questions given context text. This new system just does a better job of it. I think that what Microsoft has now is a very good first step in a new kind of technology. I wrote a demo 2 years ago that used SPARQL queries on public Knowledge Graphs to get query context text and then I used a pre trained QA model. I surprised that it worked OK. What Microsoft has is similar, but works on a grander scale. Anyway, get in the waiting list, then try it yourself.