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' capabili

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...