Except you ignored my second point which is the browser is in use throughout the system anyway. Yes it would be less likely to attack but still possible. Also a 100% safe browser does not and will never happen.
I didn't ignore it, but perhaps I wasn't clear. My point is that if Safari is in use throughout the system, but used, say, 50% less often (vs. a hypothetical 100% secure browser), then 50% of the time attacks that were previously feasible become infeasible. Frequency / probability of usage matters, not just whether something is installed.
And yes, a 100% safe browser is a fiction (thus "imagine" and "hypothetical"). But I think it's possible that another browser could be vulnerable to fewer attacks than Safari.
I think we're in agreement: adding a new browser increases the number of vulnerabilities that could potentially be exploited (assuming the new browser isn't 100% secure, which will always be the case in practice). And at the same time, moving some usage to a more secure browser makes attacks less likely in practice. I think the latter is more important than the former.
