found an interesting article talking about the real environmental impacts of AI, focused on the use/inference metrics (80-90% of total energy consumption), rather then just the model training process it also talks about how the three environmental issues of carbon, water, and land, do not move symmetrically and an improvement in the environmental footprint of one can worse that of the other issues. the total projected use of carbon, water, and land for AI data centers is projected to reach 945 Terawatt hours Wh of electricity, 9.3 trillion litres of water, and over 14,500 square kilometres of land by 2030, putting the global use of the technology above that of most countries. well the report does not go far enough on the implications of this in my opinion, it provides a pretty stark look at where the AI industry is aiming in the next 5 years and underscores how important it is for us to stop this genocidal and ecocidal technology in its tracks https://unu.edu/inweh/news/environmental-cost-of-AIs-Enrgy-use-carbon-water-and-land-footprints