That seems to start from the assumption that everyone would commit crime if they could.
In any case, the parent's point seemed pretty empty.
> Who's to say that the current deterred subset of the population won't begin to commit crimes if our style of punishment is changed?
This is what research is for. We may never know what is optimal in an absolute sense, but we can figure out some relative costs/benefits of different approaches to problems and choose our tradeoffs appropriately.