Hacker News
new
past
comments
ask
show
jobs
points
by
derektank
3 hours ago
|
comments
by
jcgrillo
2 hours ago
|
[-]
Nor is the LLM output tightly coupled to the training data. They'll "eagerly"[1] fill in the blanks wherever it sounds good.
[1] here I don't mean to imply agency, just vigor.
reply