Hacker News new | ask | show | jobs
by oefrha 1199 days ago
In the pre-LLM world we already had rumors/myths/lies/propaganda becoming "facts" after being reinforced through repetition without verification. (Relevant xkcd: Citogenesis https://xkcd.com/978/.)

Now, instead of authoritative sounding humans who are possibly wrong or with an agenda, we have an "infallible", "impartial" oracle capable of inventing whatever it wants to. Yes, I have seen ChatGPT treated as infallible on this very forum: "you are wrong because ChatGPT says <insert nonsense here>". Can't wait to see LLMs taking "post-truth" to a whole new level. The propaganda potential is immense, to point out just one application.
2 comments
theonlybutlet 1199 days ago
Perhaps it's a good thing, everyone will have to become experts in discerning the truth. Much like Video Game graphics looked "real life" until the next gen came out, perhaps society will start to see through any rumours/myths/lies/propaganda. Either that or we descend into a state of anomie.
link
oefrha 1199 days ago
Doubt it. I can hardly verify things happening in my city even if I try, let alone e.g. things happening on the other side of the globe which the media keeps telling me about in order to drum up the next cold/hot war. I used to be able to tell Photoshop jobs by non-experts, but soon enough everyone and their mum will probably be able to produce completely fake videos indistinguishable from real ones, so even "video evidence" will mean little.
link
brewtide 1199 days ago
100% on how it "takes over the world". We treat it like an oracle, hand it over willingly in mindset alone.
link