I just saw someone describe training machine learning models as "rewardmaxxing" and - while not wrong - yikes … but if training a model is “rewardmaxxing,” then I guess overfitting to training data is “rewardgooning”