A made post about this with some takeaways. TLDR/opinionated: The paper is a strong start but has some substantial limitations and needs more caveating. 
 
I realize the paper is already a few years old. Would like to see and evaluate more recent work in this area. 

Naturally this paper is several years old. My own take" we need more work in this area...  perhaps follow-up work doing a similar survey, taking sample selection and question design more seriously.

I hope we can identify & evaluate such work in a timely fashion.

E.g., there is some overlap with

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5021463

which "focuses on measures to mitigate systemic risks associated with general-purpose AI models, rather than addressing the AGI scenario considered in this paper"

You mean it will tend to 'choose' higher valence things? That would seem to make sense for biological systems perhaps, as the feelings and valence would evolve as a reinforcement mechanism to motivate choices that increase fitness.

But I'm not sure why we'd expect it to evolve similarly in a construction like a deep learning AI. No one coded valence in, and no one would really know how to access it even if they wanted to code it in, since we don't really understand where consciousness comes from.

If something in these models are sentient in any way, and if their experience have valence, I don't think we should expect "asking the chat tool what it likes" to be informative of this.

(My thoughts on this are largely the same as when I wrote this short form.) 

Btw the "whom do you hope to convince" was a sincere question, not a rhetorical question. I expect that some people are convinceable even if the change is not immediately seen.

Thinking of things like unjournal evaluation packages (unjournal.pubpub.org) I generally linkpost rather than crosspost because

• having different versions out there of the same content can be confusing,
• one might be updated not the others,
• the conversation at each location may overlap
• I'm reluctant to make very long posts with a great deal of technical content, cluttering up the EA forum in some way.

It’s generally never an issue of permission or access to the content.

but am I thinking about this wrong? Should I do more cross posting?

well written and I agree with most of the points. But I’m not sure who you hope to convince with this and how you expect to present it to people? Do you have any feedback on that? What have you tried?

Some of the points seem particularly likely to antagonise people who are Trump supporters but have some doubts for example "the idea that we shouldn’t care at all about foreigners dying by the millions is deeply wicked, and one who seriously adopts it has lost part of their soul."

I suspect that when you pose things as taxes, they almost always come out unpopular in polls. If you wanted public support you could instead lead with plant subsidy or plant based protein subsidies and then note that these would be financed by a small tax on harms to animal welfare or something.

In fact probably would do better by framing it in terms of taxing “animal-hours spent in cages”, animal pain, or factory farming in general

I added some quick polls to this post. I'd love to get some feedback to get a sense of whether we're on the right track and foster discussion. I think the polls also highlight some of the key issues and implications.

https://forum.effectivealtruism.org/posts/3Eh8MbqLwFBsD7GK2/how-much-do-plant-based-products-substitute-for-animal

State a probability for the underlying policy question:

What's the probability that investing in improving and promoting plant-based ~meat is the most effective farmed animal welfare intervention (at current margins)?

Rate your agreement over the plausibility of existing methods:

Survey and hypothetical choice-experiments are useful for helping us understand whether plant-based meat substitutes for animal products

Demand estimation from supermarket data is useful for helping us understand whether plant-based meat substitutes for animal products

Comparing this to another prominent approach (rate your agreement)

Plant-based meat is more promising than cell-cultured meat for animal welfare

I added some polls (at top and integrated in relevant sections) to get a pulse-check and foster discussion.