Musk is prone to speak in a somewhat hyperbolic fashion, and may well be more risk-averse when it actually comes time to actually decide whether to put human lives at risk, than he is when talking about that decision as an abstract future. It also isn't entirely up to him–it is also what the rest of the SpaceX executive team feels comfortable with (especially Gwynne Shotwell), and also the comfort level of the regulators and lawyers.
Even NASA "doesn't mind risking/losing some lives", in that while they drive the "probability of loss of crew" as low as feasible, they never can get it to zero. Possibly, a private SpaceX mission might have a higher go/no-go threshold for that probability – but if it adds up to (say) 50%, I really doubt they'll go ahead with the mission, they'll likely instead delay so they can invest further engineering resources in reducing it to a more reasonable level.
Even NASA "doesn't mind risking/losing some lives", in that while they drive the "probability of loss of crew" as low as feasible, they never can get it to zero. Possibly, a private SpaceX mission might have a higher go/no-go threshold for that probability – but if it adds up to (say) 50%, I really doubt they'll go ahead with the mission, they'll likely instead delay so they can invest further engineering resources in reducing it to a more reasonable level.