If people thought that your suggestion really would save tens of thousands of lives, then speed limits would probably be lowered. But your belief is just flawed.
No, he's right. People accept that tens of thousands of people are going to die every year as a direct, forseen consequence of how our transport infrastructure is designed. If we went back to a horse-and-buggy system, tens of thousands of innocent lives would be saved each year. The problem is that everyone would be condemned to a miserable standard of living, and many more innocent lives would probably be lost as a result. Seemingly prosaic things like transport and land-use policies have profound human consequences in the context of an industrialized nation-state (particularly one as massive as the United States), and this renders simple maxims like "loss of innocent life is unacceptable" quaintly irrelevant. Hopefully technology will eventually alleviate some of these hard trade-offs, but in the meantime we're stuck with them.
You have a strange definition of "accidental". The death penalty is deliberately inflicted on a person. Nobody intends a car accident.
Not on an
innocent person. Nobody intends to execute people for crimes they did not, in fact commit, just as nobody intends to crash their car. That both outcomes do end up occurring is a side effect of a system designed to produce other benefits (transport in the case of cars, deterrence/punishment/justice in the case of the penal system). Are you intentionally being obtuse here?
But let's assume you're right, for the sake of argument. Then what follows? I still don't see what you're driving at.
Quite obviously he's making the point that, just like the transport system, the death penalty needs to be evaluated by weighing the costs against the benefits. That some number of innocent people will end up getting executed does not, by itself, make the death penalty a bad idea. You have to also answer questions like "how many innocent lives will be saved by the deterrent effect?" You are of course welcome to disagree with any particular assessment of the cost-benefit analysis (and the one presented by Baron is not particularly compelling in either its construction or support), but acting baffled by the obvious fact that
any justice system is associated with some loss of innocent lives doesn't add up to a convincing argument.
Perhaps this is one of the deleterious side effects of living in a country where the issue is long settled: you are not exposed to serious debate or opposing views growing up, and so get the idea that no controversy exists, leaving you ill-prepared for encounters with people who have other ideas on the matter.
Also, I very much doubt your earlier assertion that resources are disproportionately put into preventing traffic accidents, but not death penalty accidents. On a per-capita basis (number of auto passengers on the one hand, and number of capitol cases on the other), I'd be very much surprised if the dollars-per-life weren't higher in the case of capitol prosecutions. But, anyway, that would only be an argument for spending more on death penalty cases, not for doing away with them entirely.
