First, I want to emphasise that the curves are highly contestable. My goal isn’t to argue that these particular curves are correct, but rather to encourage readers to think explicitly about what their own curves would look like.
That said, here are my thoughts on your specific points:
Effort of vegan vs carnivore diets: My intuition is that a carnivore diet can often be implemented largely by leaving out certain plant-based foods, whereas a vegan diet usually requires suitable replacements to maintain a nutritionally complete diet. But I can also see arguments in the other direction, especially given that veganism is much more common and socially accepted than carnivore diets in many contexts. What I’m much more confident about is that, for most people currently eating a typical omnivorous diet, eating more animal-based foods is generally easier than eating fewer.
Health risk of vegan vs predominantly animal-based diets: As explained in the assumptions section, I’m comparing reasonably well-planned implementations (better than average, but not theoretically optimal). So “predominantly animal-based” shouldn’t be interpreted as a diet centred on red and processed meat (although I think the evidence against unprocessed red meat is relatively weak). My own best judgement is that a well-planned predominantly animal-based diet carries lower health risk than a well-planned vegan diet, but this is certainly contestable.
Wild animal welfare beyond climate change: In principle, the external harm curve is intended to reflect all harms. However, soil animal welfare is a domain with substantial empirical uncertainty. As a result, I currently don’t have sufficiently confident views to incorporate those considerations into the curves, but they could certainly matter under some assumptions.
Thanks, these are all good points.
First, I want to emphasise that the curves are highly contestable. My goal isn’t to argue that these particular curves are correct, but rather to encourage readers to think explicitly about what their own curves would look like.
That said, here are my thoughts on your specific points:
Effort of vegan vs carnivore diets: My intuition is that a carnivore diet can often be implemented largely by leaving out certain plant-based foods, whereas a vegan diet usually requires suitable replacements to maintain a nutritionally complete diet. But I can also see arguments in the other direction, especially given that veganism is much more common and socially accepted than carnivore diets in many contexts. What I’m much more confident about is that, for most people currently eating a typical omnivorous diet, eating more animal-based foods is generally easier than eating fewer.
Health risk of vegan vs predominantly animal-based diets: As explained in the assumptions section, I’m comparing reasonably well-planned implementations (better than average, but not theoretically optimal). So “predominantly animal-based” shouldn’t be interpreted as a diet centred on red and processed meat (although I think the evidence against unprocessed red meat is relatively weak). My own best judgement is that a well-planned predominantly animal-based diet carries lower health risk than a well-planned vegan diet, but this is certainly contestable.
Wild animal welfare beyond climate change: In principle, the external harm curve is intended to reflect all harms. However, soil animal welfare is a domain with substantial empirical uncertainty. As a result, I currently don’t have sufficiently confident views to incorporate those considerations into the curves, but they could certainly matter under some assumptions.
Hope this is helpful!
Thanks for the response Chris