not sure if only the paid gpt actually cites, but anyone thats trying to show someone else how inefficient LLMs are with data should be able to come to the conclusion the output is untrustworthy if you’re getting 404 errors after asking to cite, in my opinion
have the AI cite it sources for the claims it makes, and read through those sources. ask the AI what information it used in the source to come up with its statement.
using chatGPT in this way is like how I used Wikipedia growing up: just pull up an article on something, check out what citations an article had, and walk it backwards from there
+1 for yunohost from me. its how I stood up my first instance for the fediverse and self hosting.
I think its very easy to comprehend and is a great way to join the fediverse on hardware you manage
same with washing hands. 2 minutes max on a Sunday is enough for the rest of the week
the amount goes down a bit when the lid is closed though, no?
same, I don’t need farticles spewing everywhere
I wouldnt take anything I see on the internet seriously, especially in a community like one that has the name “shitpost” in it, lol
what’s a lib
same. only a matter of time till it happens here too
you’re gonna have to conduct some research for us, lmk how that goes!
artfully stylized graphics will always be better because they will look timeless
probably bill clinton links to epstien
what did you eat today
all I wanna do is 🔫🔫🔫🔫
I guess we found the answer to how the original person i was replying to can validate anything thats generated by an LLM: looking up sources that an LLM claims to use