You could probably come up with something similar to Moore's law for for those stats over those first 50 years. Makes me wonder if/when computing will reach a similar plateau. It's felt like we're close for a while, but at least if you allow for a somewhat flexible definition, it just keeps trucking along.