It's merely revealed that any massive, AGI-focused model is actually vastly more inefficient than we had even originally suspected it would be.