Meat was useful, back when we had not yet selectively bred fantastically better than natural crops of all kinds, back when we had not yet invented synthetic fertiliser that's now the ultimate source of 70-80% of the nitrogen in the body of someone in an industrialised nation, back when hunger was a bigger problem than obesity.
Now? Now meat's mostly a problem, not a good thing. Even if you ignore every ethical argument, regardless of if your concerns are your own health or the environment, meat's not good.
Data centres… well, I think this is a bubble, I also want it to be a bubble for various reasons, but the AI running on them today is in fact already useful.
Even if current AI wasn't at all useful (despite it having about half to one quarter of the market size as meat already), it does so at a cost orders of magnitude lower environmental harm than meat. Convincing half of the population to have "meat-free Mondays" (so, reducing consumption by 1/14th) would do more than switching off all the AI DCs, given the estimates from Greenpeace for AI https://www.greenpeace.de/publikationen/20250514-greenpeace-... and Our World In Data's estimates for livestock and manure https://ourworldindata.org/ghg-emissions-by-sector
I'll be the first to cheer if we get rid of industrial agriculture but there's an awful lot of land in the world that doesn't receive enough rain for farming but which is still fine grazing land and when used for grazing still supports most of its original ecology. And there's a lot of damaged, blemished, etc produce that pigs are happy to eat but which can't be sold in a supermarket.
I'd like to see meat consumption to something like half to a quarter of its current level rather than eliminate it outright.
OK, but (1) also a lot of good land is being used to feed livestock, the biomass of livestock is quite a bit higher than the biomass of humans; and (2) even reducing it just by a quarter is several times more than the combined impact of all the AI data centres.
> Does this come from a Big AI talking points memo?
It comes from the evidence I linked you to.
Which includes, to repeat, *Greenpeace*.
Also to repeat: I think this is a bubble, I also want it to be a bubble for various reasons.
As in, I do not buy into Big AI's talking points about how this is "it", and we're on a path to radical AI-based abundance. Not yet. Plus I think it would be bad even if we were on that track at this point, so I want it to be "not it".
> ridiculous
The global meat market is around 1.5 trillion USD, give or take. That is literally the value of meat, which like all things in a free economic sector can be measured in money.
You may also notice from me saying that AI is 0.5-0.25 of that, that I'm not using "Market Cap" of AI in this comparison. Market cap != market size. This is about what revenue AI and meat gets per year.
Now? Now meat's mostly a problem, not a good thing. Even if you ignore every ethical argument, regardless of if your concerns are your own health or the environment, meat's not good.
Data centres… well, I think this is a bubble, I also want it to be a bubble for various reasons, but the AI running on them today is in fact already useful.
Even if current AI wasn't at all useful (despite it having about half to one quarter of the market size as meat already), it does so at a cost orders of magnitude lower environmental harm than meat. Convincing half of the population to have "meat-free Mondays" (so, reducing consumption by 1/14th) would do more than switching off all the AI DCs, given the estimates from Greenpeace for AI https://www.greenpeace.de/publikationen/20250514-greenpeace-... and Our World In Data's estimates for livestock and manure https://ourworldindata.org/ghg-emissions-by-sector