Thanks, very strange. I definitely selected linkpost; the GUI appears to have forgotten this, possibly when I tried to edit the thumbnail. Fixed.

Thanks, done.

Yes I find it bizarre that many people seem to be simultaneously opposed to the War on Drugs and also support extending it to a new drug. 

A final way the tobacco industry is dodging the rules is with e-cigarettes, or “vapes.” Using marketing that illegally targets children, they peddle vapes as a “safer” way to smoke. This tactic has proved alarmingly successful. 

SMA condemn the tobacco companies for claiming that vapes are safer, but don't discuss whether this key claim is actually true. Yet as far as I can see it clearly is true. There is debate about exactly how much safer they are - e.g. how convincing we should find the NHS claim that vapes are 95% safer - but I haven't seen any credible argument that vapes aren't safer at all. It's not 'dodging' safety rules to release a considerably safer product.

Further, I think vapes are also pretty good evidence again SMA's defense of paternalism. If smoking cigarettes wasn't really a choice, why has the availability of vapes and pouches been associated with a decline in cigarettes? The most natural explanation here is that previously people choose to smoke cigarettes, and then a superior product came along, so people started choosing that instead.

If you think it's merely 'arguable' that OpenAI has had had a significant negative impact through acceleration then I think you are significantly more positive than the median EA.

Thanks for sharing! TruthSocial having more positive engagement is interesting.

I agree this post is within scope, but that is because it is about AI and policy. It's not because StopAI has any nontrivial EA support.

A StopAI organizer has posted here before, and received a mixed reaction from the community.

The post got -29 karma, which is an extreme outlier for how negative it is. Unless by "mixed" you mean 'not literally everyone disliked it', I think by any reasonable account the post received a decidedly negative response. 'a guy who wrote a downvoted post has a co-organizer who did something bad' is not enough to make something notable - if our standards were that low, almost anything would be within scope.

To give one example of practical relevance, the post immediately above this one (on my current feed) considers financially supporting StopAI, although it expresses concerns about their tactics. 

I think this is an unfair summary, making the post sound significantly more positive to StopAI than it is. Michael considered donating, having decided not to in the past, and then decided to continue not donating, as he had "become more confident in [his] skepticism".

so voter preferences cannot be opposite of what is best for human welfare, by definition.

This is clearly not true. The example I gave was foreign aid, which benefits foreigners at the expense of citizens. Since only one of these groups can vote, there is little reason to think that the preferences of this subgroup will align with overall human welfare. And we know it doesn't - hence the polling data.

This is true for most EA cause areas. Existential risk work is about protecting the interests of future generations; animal welfare work is about protecting the interests of animals - neither of which groups can vote. 

the page directly addresses that question quite incisively, citing the bayesian regret figures.

No methodology or source is given for why we should expect a 5% decline in the risk of 2 billion deaths.

bayesian regret figures by princeton math phd warren smith show that approval voting roughly doubles the human welfare impact of democracy.

Their result is that, in their model, outcomes more closely match voter preferences. But my example is one where voter preferences are opposite to what many EAs think is best for human welfare. 

doing some ballpark math to see how many lives that would save:

Suppose the USA, by adopting range voting and thus making better decisions, lowers the risk of a 2-billion population crash in 50 years, by 5%. I consider this a conservative estimate.

These numbers just seem totally made up. Why should we believe that approval voting has anything like such a large impact?

I think it's better to try to keep the upvote/downvote and agree/disagree axis more distinct - to express normative vs positive evaluation. So in this case I would think that downvoting would be the natural way to express disapproval.