I notice a huge influx of substack reposts on the forum, and with it, a decline in community norms around writing style. 

It's unfortunate that blog posts are generally designed to attract readers, and rarely to seek truth.

In most worlds where AI disempowers humanity, the human species continues

Do you mean most likely worlds? The difference seems incredibly important - there are, in my view, quite compelling arguments that the most likely outcome given disempowerment is human extinction, but of course I can imagine worlds in which that doens't happen.

Right, what seemed to be missing to me was evidence that longtermists specifically had stopped working on extinction risk. But I see there are a good amount of posts others have linked in the comments that would count as evidence for this.

Is this post conflating EAs or AI-safety researchers/advocates with longtermists? My impression is that actually rather few EAs are strong longtermists, and AI safety researchers/advocates maybe even less - they're just united by their wish to avoid catastrophe, but differ in their understandings of the kind of catastrophe we should expect.

I think when AI safety was young and, er, weird, a lot of those talking about it were longtermists. That makes sense. But now it's become a mainstream EA concern, and even a not-weird concern in society generally, so it's unsurprising that it's become less longtermist, from a sociological perspective.

I quite liked this article by Martha Nussbaum: Virtue Ethics is a Misleading Category. She points out that both the classical utilitarians and Kant talked extensively about virtues. On the other hand, there's great variation among those who call themselves 'virtue ethicists', such that it's not clear if virtue ethics is really a thing.

But the point I want to make is: a good utilitarian has to acknowledge the role of virtue, and I think a lot of modern utilitarians have forgotten this. We want to use utility-calculation to guide our actions, but humans can't think like calculators all the time. 

I'm not really into deontological constraints myself. Rules of thumb, yes, but they should always be open to revision. Exceptional circumstances can always justify breaking rules - and in those cases, I will refer to what maximizes utility.

There are a few things missing from this argument, that I think are important to add:

1. Even on "humane" farms, most animals will only live a few months. It takes some clever twisting of logic to say that bringing someone into existence only to kill them a few months later isn't wrong. 

2. Even on the ideal free range farm, animals don't live as well as they could. E.g. where I grew up, there are free range sheep. But they are still treated roughly by the farmers, they still die out on the fields if they get sick or too cold. Why settle for humane when we could actually care for animals the way we do for our pets?

3. Donating is great, but political action is also essential if we ever hope to achieve fundamental change for animals. Join protests, sign petitions, write to your representatives and vote for politicians who promise to fight for animals. 
 

I'm curious actually why 80,000 hours didn't aggressively expand in the past. It seems crazy to me that one of the core EA organisations only has 50 staff after existing for 15 years. Did you not feel growth was justified before, or has it just been a gradual process that you're now pushing harder to accelerate?

I suppose it's a reaction to the tendency on the political left to not listen to a person at all because of some association they have with some group.

But I agree with you. We should be wary of these dynamics, without falling into black and white ways of thinking.

Also to second Nick, I really 'felt' and resonated with this post!

I find the difference between pain and suffering (1&2) plausible, and it also makes sense that the sense of self needed for suffering is more likely to arise in social species. It also seems plausible that this creates a greater welfare range, insofar as it adds on another type of unpleasant experience.

I'm hesitant to place too much importance on this when thinking about the badness of the experiences of simpler organisms, though. For one reason, as you pointed out, it doesn't seem to change the policies we should support most of the time. But secondly... I think it's easy to overestimate the importance of our own self-consciousness. I actually think that most of the time I'm not self-conscious - I'm just going about my day, doing things habitually. But physical pain can still be awful, for example a headache, or when I stub my toe after waking up. This is clearly different from e.g. depression, which I think plausibly does require self-consciousness. I doubt that insects can get depressed. But I feel like we can make pretty good assessments of the quality of their lives on the basis of non-reflective experience.

Those are just my thoughts, I don't think I'm disagreeing with anything in the post!

I also note that expected lifespan, while seeming like a pretty good indicator of welfare, is going to be extremely sensitive to assumptions about when an organism becomes sentient for r-selected species. E.g. whether a trout is already sentient in an egg, at the point of hatching, or only two days later might change its expected lifespan from 0, to <1 day, to maybe a week (these are guesses, but I think they're quite plausible).

I don't think that undermines the measure, it just shows how hard this is... :/