OTOH you can't get sufficient sampling on anything but the smallest systems to make any sort of dynamic prediction about it. Not to mention the approximations that were made during the choice of basis sets for the atomic orbitals and how one decides to handle the configuration interaction.

TLDR: Approximations are everywhere, make sure you work that into your conclusions.

I'll clarify that when he says sub-nanosecond timesteps, he means way sub-nanoscale timesteps. Femtosecond timesteps are used at the molecular dynamics level. Even smaller ones are used at the QM level.