I've watched Lawrence Krauss's lecture and read his book. I think I got what he was saying, and I don't have any problems with that; however, what I can't get is how the laws of physics that makes this phenomenon possible came to be in the first place?
What we call "laws of physics" have an evolutionary path.
It really started with Newton and the falling apple and slowly it evolved into complete mathematical models of experimental observations, called theories. From the observations and the theories conservation laws emerged.
These laws are strictly obeyed within the framework that they have been validated. Take the thermodynamic law "entropy remains the same or increases in closed systems" . The region of validity of the law was transformed when the atomic nature of matter became understood, and statistical mechanics became the underlying framework. There, from an absolute law it became an estimate of probability outcomes, which to all intents and purposes recreates the law for macroscopic systems.
As our observations and experiments advance, new mathematical frameworks appear which transmute the laws of the overlying frameworks : conservation of energy became the relativistic four vector energy which blended mass in the mixture. Conservation of relativistically defined energy and momentum also became fuzzy instead of absolute due to the Heisenberg uncertainty principle of Quantum Mechanics .
So when we come to cosmology where there exist theoretical models of solutions of General Relativity it is not surprising that apparent inconsistencies with conservation laws developed for different frameworks.
At the moment there does not exist a theory of everything which quantizes gravity and includes the other three forces, weak, strong, electromagnetic that has been validated through all relevant observations, even though string theory offers such possibilities. It is therefore premature to be definitive of how the known conservation laws validated by our laboratory experiments will evolve in a cosmological setting. Something and nothing have to be mathematically defined within the appropriate theoretical models.