If the laws of AI scaling continue to be true, does this mean that in a year or two we would be seeing frequent revolutions in mathematics and physics?