Why is it better to have more implementations than all focusing on one... Seems to me this would be a way to introduce more vulnerabilities and not less.
The assumption is that there are diminishing returns to trying to make one code base secure. Eventually you'll have to spend twice the effort to make it a little bit more secure. At that point multiple implementations make sense because as long as two implementations don't have the same/compatible vulnerabilities there can't be loss of funds if only one is compromised.
1
