News for hackers. Stuff that matters.

A comment on GPT-4 performs significantly worse on coding problems not in its training data. Parent comment

Re:GPT-4 performs significantly worse on coding problems not... by saulpw on Saturday March 25, 2023 @12:56AM (#35299587)
To be fair, using APIs that I think should exist, is how I develop most of my APIs.
[ Parent | on HN ]
Re:GPT-4 performs significantly worse on coding problems not... by jasfi on Saturday March 25, 2023 @01:24AM (#35299733)
Except that I wasn't asking it to develop a new API.
[ Parent | on HN ]
Re:GPT-4 performs significantly worse on coding problems not... by saulpw on Saturday March 25, 2023 @01:30AM (#35299768)
Then I guess you're not using it to its fullest potential ;)
[ Parent | on HN ]
Re:GPT-4 performs significantly worse on coding problems not... by bamboozled on Saturday March 25, 2023 @04:33AM (#35300661)
You can’t keep blaming the user.
[ Parent | on HN ]
Re:GPT-4 performs significantly worse on coding problems not... by brabel on Saturday March 25, 2023 @04:40AM (#35300709)
It's very likely it was using other languages' as "inspiration" given there's very little Zig code out there... so it's maybe natural it would use APIs that don't yet exist... perhaps informing it that it also needs to implement those APIs could work?
[ Parent | on HN ]