Y
Hacker News
new
|
ask
|
show
|
jobs
by
anktor
1166 days ago
Isn't the paper discussing precisely the opposite? That chatgpt predicts text in a way that, with each version, resembles more and more human logic, both with its succeseses and errors?