AIA historical bug in LLM scaling laws led the AI industry to train oversized, undertrained modelsGoogle DeepMind's Sander Dieleman highlighted the costly pre-Chinchilla compute errordigg it1800123456789012345678901234567892 Sources, 140d ago, first seen 140d agoCombined views94.5K2 Sources, first seen 140d ago625 likesCombined views94.5K2 Sources, first seen 140d ago625 likes