Re:Bing sees 15.8% boost in visits following GPT-4 push, as ...
by AviationAtom on Friday March 24, 2023 @02:56PM
(#35293841)
I love that Bing Chat actually links to references. It felt like something was missing from ChatGPT without it.
[ Parent
| on HN ]
Re:Bing sees 15.8% boost in visits following GPT-4 push, as ...
by HeavyFeather on Friday March 24, 2023 @03:09PM
(#35294006)
Just because it does, it doesn’t mean they’re being interpreted correctly, so watch out for https://xkcd.com/906/
Re:Bing sees 15.8% boost in visits following GPT-4 push, as ...
by AviationAtom on Friday March 24, 2023 @03:34PM
(#35294330)
In the context of LLMs they aren't really sources, just additional reference material. I'd trust what Bing Chat tells me far before I'd trust Google's new Bard AI. It legit started making up non-existent commands for me to use.
Re:Bing sees 15.8% boost in visits following GPT-4 push, as ...
by rtsil on Friday March 24, 2023 @04:30PM
(#35294999)
If I have to check each reference, what's the added value compared to just giving me a list of links (i.e. a good old search engine)?
Chat search tools are only useful if (when?) they achieve a 99.99% reliability (I'm fine with one mistake per 1000 searches).