|
Matias Koivurova@mkoivurova.bsky.social |
This is the most in-depth article on bullshit that I've ever read. And I feel that it is spot on 🎯
"These models aren’t designed to transmit information, so we shouldn’t be too surprised when their assertions turn out to be false."
link.springer.com/article/10.1...
4 replies 32 reposts 118 likes
|
@zendarva.bsky.social
[ View ] |
So, the thing that has no concept of truth (Or concepts for that matter) has no concept of truth. They told us this when they released them. Why do we need a study to discover what the developers told us about llms?
1 replies 0 reposts 0 likes
|
darko barsoomian
@darkoverlordofdata.bsky.social
[ View ] |
I disagree. On Bullshit implies intentional manipulation. I think AI is not intentional, but more like senility.
2 replies 0 reposts 0 likes
|
PolyChrome
@jenstone.bsky.social
[ View ] |
📌
0 replies 0 reposts 0 likes
|
Dr M
@ivalaine.bsky.social
[ View ] |
I worry how AIs fed crap data and so spit out crap will end up affecting how people feel about machine learning done with good clean verified datasets.
0 replies 0 reposts 0 likes