I would also add to point 3: don't go "whale fishing" and search in ways that generates a ton of "bycatch" (excuse the fishing analogies, it's too useful to not use for this point but irl leave the fish alone). Putting out huge calls saying "we need all hands on deck, we're heavily talent constrained and can't find people, you should apply even if you don't think you're a good fit" is going to reach a ton of people who really are not a good fit, who will think this is a worthwhile pivot because they are talented and the field is talent constrained, and waste a bunch of time applying and considering ways to orient their life to the new direction. Then when they inevitably are not a good fit, and talent constrained means something along the lines of "we're having trouble hiring the top 99.9% talent" and not "we're having trouble hiring the top 80% talent", they will get more bitter and cynical about the movement and doing good effectively, and may even miss opportunities that would be a good fit.

Research on AI suffering has higher marginal value than research on AI consciousness⁷

Depends on what you mean by the term "consciousness"

Non-human suffering is neglected by the AI safety community⁵

As I understand, many key players are explicitly anthropocentrists. Hell, some key players care more about money than humans, let alone non-humans.

Understanding how LLMs learn (e.g. SLT) is underrated relative to work on instilling specific behaviors

Understanding how they learn is how we get insights into methods of teaching them good values

We are making good progress in the AI S-risk space and research is on track

Afaik there isn't any robust research on how to use AI to end factory farming. That alone is a sign to me that we are far behind.

Theories of consciousness will lead to actionable understanding of AI consciousness²

Global workspace theory makes anthropic's discovery of the "j-space" much more plausibly a sign of AI consciousness. If a theory of consciousness doesn't help us determine between whether AI is conscious or not, I don't think it's much better than a theory of phlogiston is to fire. 

The backfire risks of AI values alignment outweigh the expected positives

If we use ai to bring factory farming to the stars, I think that will likely be worse than all the benefits it'll bring.

Benchmarks will become useless due to eval awareness¹

I think they will become over saturated and disconnected from real world usage such that they will be pretty useless, but I don't think eval awareness is what will make them lose their value

You guys get to be rational actors??

Here's an idea for a pro-animal campaign that could absorb a lot of money: subsidize vegan food. Make mock meats and other plant based stuff way cheaper, so people are more likely to try it and get used to it and like it.