Mythos is a big deal because of its second-order effects.
If it only cost $10b to train a cyber weapon that can easily penetrate all heavily-used operating systems, I expect that the information security industry will encounter many private variants within 24 months.