Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
IIAOPSW
on June 22, 2023
|
parent
|
context
|
favorite
| on:
People paid to train AI are outsourcing their work...
Try pointing out its mistakes when it hasn't made a mistake. You'll usually get the same "I'm sorry, here's the correct answer" output.
eterm
on June 22, 2023
[–]
And getting it to derive it's mistake doesn't always work (NB: This is GPT3 & bard, GPT4 is much better):
https://richardcocks.github.io/prime.html
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
