HN2
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
kybernetikos
on March 22, 2023
|
parent
|
context
|
favorite
| on:
GPT-4 and professional benchmarks: the wrong answe...
Going from crazy nonsense to wrong, but arguably human level performance (80% of humans are bad at this task), is still a nice improvement. I'll have to give it some of my logic tests and see how it does.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: