From a strictly utilitarian perspective, would there be anything to be gained by, say, starting a campaign of assassination against executives of tobacco companies?

Well, it would be unlikely to convince people not to smoke. It would create a kooky outgroup faction, the Violent Anti-Smokers, that people would be afraid to be associated with, so people would become uncomfortable with being too vocally anti-smoking lest they be mistaken for one of Them. No such campaign is likely to be effective enough to make people unwilling to become the executive of a tobacco company to the point that positions aren't filled, they'd just ramp up security.

Also, deaths-caused isn't really the best metric for toll on human society. Cigarettes rarely do anyone in at thirty, while car accidents don't discriminate. Adjusting to life-years-lost would be a step in the right direction.

New proposed censorship policy:

Any post or comment which advocates or 'asks about' violence against sufficiently identifiable real people (as opposed to aliens or people on trolley tracks) may be censored.

Reason: Talking about violence makes it more probable, makes LW look bad, and numerous message boards across the Earth censor discussion of various subtypes of proposed criminal activity without anything bad happening to them.

More generally: Posts or comments advocating or 'asking about' violation of laws that are actually enforced against middle-class people (e.g., not drug laws) may be censored on the grounds that it makes LW look bad and that anyone talking about a proposed crime on the Internet fails forever as a master criminal (i.e., even if a proposed conspiratorial crime were hypothetically good, there would still be net negative expected utility from talking about it on the Internet, therefore and in full generality this is a low-value form of discussion).

This is not a poll, but I am asking in advance if anyone has non-obvious consequences they want to point out. In other words, the form of this discussion is not 'Do you like this?' - you have a different utility function from people who are responsible for how LW looks as a whole - but rather, 'Are there any predictable consequences we didn't think of that you would like to point out, and possibly bet on with us if there's a good way to settle the bet?'

From a strictly utilitarian perspective, would there be anything to be gained by, say, starting a campaign of assassination against executives of tobacco companies?

No, that would be bad. The heuristic that says violence is always a bad strategy is reliable, and it works for reasons beyond the obvious. One important reason is that whenever someone uses violence, other people respond by trying to figure out what his goals are and undermining them. So the practical consequence would be to move tobacco into the "good" category in some peoples' minds, and to undermine anti-tobacco strategies that actually work (taxes, advertising, and regulations).

On the other hand, displacing cigarettes with nicotine vaporizers or with some other drug looks like a straightforward win.

Any citations on the numbers? I assume those aren't just the number of smokers who died from lung cancer or something as silly as that. Did you factor in the fact that air pollution kills a lot of people and is also caused by people (and cars as in your example) in different ways, for example by being one of the causes for Cardiovascular diseases(so is smoking, yes) which are the most frequent cause of death? Does pollution kill less people than tobacco? I mean you are saying that it isn't even close so you probably have, but I would not take such statements at face value without more data.

From a strictly utilitarian perspective, would there be anything to be gained by, say, starting a campaign of assassination against executives of tobacco companies?

Yea these kinds of questions resonate badly with casual visitors, and a few affirmative answers can be easily misused to paint LW in a bad light. Maybe choose another, functionally equivalent but less controversial example?

If one wants to be precise, one should measure deadly harm not in deaths (past people were still going to die if lung cancer didn't get them), but in quality-adjusted life years, pretty much the best unit ever.

And no, I'm not sure assassinating CEOs would help, unless it was public enough and effective enough to actually dissuade people from leading the company in any capacity. Bombing factories would be much more efficient.