I thought this market estimate of 68% was very off-base, in presuming quick advanced in lab-grown meat. I think doing that affordably and at scale is unlikely to happen, even over that time horizon.

I think it's actually more likely that animal farming rises by 2068 than falls, due to rising living standards in developing countries.

@MaximLott *58%

Does this include animals kept primarily for companionship or recreation or preservation?

Does this include digital animals, eg uploads, simulations, etc?

Does this include humans?

@MartinRandall
- just the food industry (meat, dairy, eggs, etc.)
- no digital beings sensu stricto
- Humans are animals, so if they end up being farmed, that population would count.

There’s just no way AI won’t have developed cheap lab meat by 2068

What an evil wish
Fake goo, in your dish
You wanted vegan cheese
But settled for soy sleaze

Africa will have something like 3 billion people by then, and I consider it highly unlikely that Africa will reach US/Western Europe levels of development by then, or have sufficiently WEIRD moral intuitions to care much about farm animal suffering.

@MrMayhem Yup. It is sadly quite unlikely to me that cellular ag and alternative proteins will make a significant dent in global trends in BRIC countries and in Africa.

@NicoDelon I can actually see China going for it iff it becomes cheaper than farmed meat. But yeah, I don't see India going for it for religious reasons - they revere the cow religiously and for them, I assume the more cows the better.

This is globally or only in the US?

@Odoacre The original question doesn't mention any particular country, so I think global is the only reasonable answer to that.

How is size measured here @JacyAnthis? Number of animals killed? On farms at any given time? Land use? Revenue?

@NicoDelon animals on farms (I had in mind the average across the year of 2068, but the question here says "by the beginning of 2068," and it shouldn't make a big difference)

@JacyAnthis Including aquatic farms?

@NicoDelon yes, as long as the animals are essentially confined in space (e.g., a net separating farmed fish from the sea even though it's the same water molecules; a grassy field with cows grazing even though they could wander off). Wild-caught fish, which is a very large category, does not count here.

@JacyAnthis But invertebrates count, and since we’re counting heads not tonnes, it’s NO!

@NicoDelon yes, lots of invertebrates.

I should also say that, personally, my prediction here is very tied up in my beliefs about AGI. E.g., with the power of an aligned, well-behaved AGI, it's hard to imagine humanity not switching to cultured meat or becoming digital minds. On the other hand, if the population of human descendants is orders of magnitude larger than it is today, then just a small fraction of humans choosing to keep farming animals could prevent this from resolving NO.

It's quite a messy thing to predict if you think the world will be radically different by 2068!

@JacyAnthis You mean ‘prevent this from resolving YES’?

@JacyAnthis Re AGI it’s also not hard to imagine it could be harnessed to accelerate the extraction of animal protein through farming. A whole lot turns on 1. What well-behaved means and 2. IF it’s likely to be well-behaved.

@JacyAnthis "becoming digital minds" - the overwhelming majority of humans will resist this vigorously, many of them violently (I am one of them - I will be uploaded over my dead body, dead in a way that will make uploading impossible). I do not see this as a factor unless there's a major alignment failure almost bad enough to cause a X-risk or S-risk. EA is the most extreme manifestation of WEIRD values (in the moral psychology sense of WEIRD), and 90% of humans don't share WEIRD values to begin with. And that's not getting into the many, many people in WEIRD countries who are religious and/or like physical existence. Even a very strong majority of PhilPapers survey respondents said they would not enter the experience machine (probably analogous enough to mind-uploading to be useful for this argument), and judging by the answers to most of the other questions in that survey, it's pulling from an extremely WEIRD sample.

Related:

@IsaacKing yeah I'd put that at 5-6%, with most of that probability-space covering widespread and intense coercion. If you condition on no widespread coercion, probably <2%.

Cultured meat, on the other hand, is much easier to envision - all it takes is governments regulating it out of existence. I consider this unlikely on a global scale, but orders of magnitude more likely than most humans choosing to upload their minds.