#computing #ai # [[Epistemic status]] #shower-thought # Related - [[Metaphors are data augmentation]] - [[Train young AI with grown AI]] - [[Open-endedness]] # Dataset augmentation You can augment language dataset for example a lot with [[GPT3]], I wonder if there is research on the capacity to augment a dataset, the limits. Because I’m not sure, [[Does new information actually exist]]? Or after augmenting a dataset by 900% it will only generate in the [[Gaussian]] ?