Agreed. The number of people affected is a big difference. But I think the comparison does make clear that isn’t necessarily a principled reason not to accept a few failures as part of otherwise useful regulation.
The context in this discussion is comparing the hypothetical of a homeless person dying due to a systemic failure to the hypothetical of a business going bankrupt (without being acquired) due to a systemic failure.
This concept of presumed innocence ought to apply to this hypothetical homeless person: why not give them shelter (analogous to allowing M&A) by default, until someone decides it’s a problem? If a society is willing to accept this personal death, they should be willing to accept a business failure.