congrats on this very well written, interesting and useful article. Before starting, I'd never thought I'd finish it, but very glad I did. Only time i got sidetracked was when i took your advice and started a convo with Claudi on sentience, which took longer than expected :)
I think this might be the longest article I've ever read, and while I'm still processing it all, I'm quite certain that my approach to many parts of my work is going to change significantly (and hopefully become more effective), as a result.
Thanks a lot for the time and effort this must have taken to write!
Included in our newsletter (will be published today) because I respect Joseph's opinion, and he was impressed with this! I appreciate the audio version - have saved it to my podcasts! - Sam
Also, I'd encourage you to linkpost this on the EA Forum. While there's a lot of discussion there, as you highlighted, I'm sure many users would be excited to know about this post!
The AlphaMeat analogy is spot on, optimization at that scale could finally crack the cost and taste problems that have kept BYND and others from mass adoption. Plant-based proteinn development right now is basically trial and error with limited resources, but AI could run millions of formulation experiments in silico before a single lab test. The cultural resistance piece is what worries me though, because even if we solve the technical challenges people get weirdly emotional about defending their meat consumption. Your caveman comparison is fair but the exponential nature of AI progress makes me think we have a narrower window than you're suggesting to get this right.
I agree on the cultural resistance to clean meat point (actually wrote about that back when I was at Pax Fauna: https://paxfauna.org/social-norms-blind-spot/) though I’ve updated to be less worried about cultural resistance given cultured meat will be one of the least weird things I expect to happen in the coming decades. As for narrow windows of opportunity, I also agree there. It’s such a hard balance to strike because over adjusting for a timeline that turns out to be faster than reality could be just as harmful as moving too slowly.
After several hundreds of hours discussing this topic and writing about it, I still doubt that "anticipating" will be better for animals in expectation. However, this is might be the first "intro post from first principles" about this, and it goes pretty deep into the topic, so I admit I'm impressed!
Sadly, again, I do think our situation when figuring out how to defend animals in a post-AGI world is very similar to that of a caveman who'd wish to take action during his lifetime in -12,000 BC to effectively improve the treatment of animals in the 21st century. But I think you highlight this quite well, and that further exploration of the topic makes sense on principle.
I've become less optimistic about anticipation-type interventions probably for the same reasons, but not enough that I'm ready to give up thinking about them, though I probably haven't spent as much time sorting out my thoughts. Two important disanalogies with cavemen are 1) we have the situational awareness to know dramatic change is coming, i.e. we are even capable of asking the question, and 2) we have a lot more past data to look at about what OOMs of economic growth look like. I think we shouldn't count ourselves out! We're definitely in an unprecedented position but we can reflect on that position in an unprecedented way.
congrats on this very well written, interesting and useful article. Before starting, I'd never thought I'd finish it, but very glad I did. Only time i got sidetracked was when i took your advice and started a convo with Claudi on sentience, which took longer than expected :)
That’s exactly the reaction I was hoping for, and you’re exactly the kind of person I was hoping to hear it from!
I think this might be the longest article I've ever read, and while I'm still processing it all, I'm quite certain that my approach to many parts of my work is going to change significantly (and hopefully become more effective), as a result.
Thanks a lot for the time and effort this must have taken to write!
You just made my week, Mel! Even one response like that makes all the time well worth it!
Included in our newsletter (will be published today) because I respect Joseph's opinion, and he was impressed with this! I appreciate the audio version - have saved it to my podcasts! - Sam
Wooo thanks Sam & team!
Also, I'd encourage you to linkpost this on the EA Forum. While there's a lot of discussion there, as you highlighted, I'm sure many users would be excited to know about this post!
Thanks for the push, I'll share it there!
The AlphaMeat analogy is spot on, optimization at that scale could finally crack the cost and taste problems that have kept BYND and others from mass adoption. Plant-based proteinn development right now is basically trial and error with limited resources, but AI could run millions of formulation experiments in silico before a single lab test. The cultural resistance piece is what worries me though, because even if we solve the technical challenges people get weirdly emotional about defending their meat consumption. Your caveman comparison is fair but the exponential nature of AI progress makes me think we have a narrower window than you're suggesting to get this right.
I agree on the cultural resistance to clean meat point (actually wrote about that back when I was at Pax Fauna: https://paxfauna.org/social-norms-blind-spot/) though I’ve updated to be less worried about cultural resistance given cultured meat will be one of the least weird things I expect to happen in the coming decades. As for narrow windows of opportunity, I also agree there. It’s such a hard balance to strike because over adjusting for a timeline that turns out to be faster than reality could be just as harmful as moving too slowly.
After several hundreds of hours discussing this topic and writing about it, I still doubt that "anticipating" will be better for animals in expectation. However, this is might be the first "intro post from first principles" about this, and it goes pretty deep into the topic, so I admit I'm impressed!
Sadly, again, I do think our situation when figuring out how to defend animals in a post-AGI world is very similar to that of a caveman who'd wish to take action during his lifetime in -12,000 BC to effectively improve the treatment of animals in the 21st century. But I think you highlight this quite well, and that further exploration of the topic makes sense on principle.
I've become less optimistic about anticipation-type interventions probably for the same reasons, but not enough that I'm ready to give up thinking about them, though I probably haven't spent as much time sorting out my thoughts. Two important disanalogies with cavemen are 1) we have the situational awareness to know dramatic change is coming, i.e. we are even capable of asking the question, and 2) we have a lot more past data to look at about what OOMs of economic growth look like. I think we shouldn't count ourselves out! We're definitely in an unprecedented position but we can reflect on that position in an unprecedented way.
Agree. We need to stop relying on the current leading orgs to make change.
I agree that small, agile, and likely new orgs will generally be better positioned to respond to the dizzying changes headed our way!