Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
parasubvert
on March 20, 2023
|
parent
|
context
|
favorite
| on:
GPT-4
I'm not sure I care about how it does the work, I think the interesting bit is that the model doesn't know when it is bullshitting, or the degree to which it is bullshitting.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: