Re:Github Copilot Wants to Play Chess Instead of Code
by cjbgkagh on Thursday January 13, 2022 @11:09AM
(#29922283)
It would be nice if it could detect and avoid tautologies; more difficult would be to avoid circular reasoning but at least it should be possible to make the circles bigger.
[ Parent
| on HN ]
Re:Github Copilot Wants to Play Chess Instead of Code
by notahacker on Thursday January 13, 2022 @11:33AM
(#29922724)
I was actually surprised by the sentence "bad means something is bad" given that substituting the second "bad" for a synonym is well within NLP models' capability, and you'd expect training processes to eliminate that level of tautology in model outputs.
Then I remembered Github copilot wasn't optimised for natural language, it was optimised for programming, where [near] tautology isn't unusual and variable names aren't supposed to be swapped out...