The energy wastage here is exactly what I was talking about to many AI bros here who continue to consume tons of water from rivers to be used for inefficient tr

The energy wastage here is exactly what I was talking about to many AI bros here who continue to consume tons of water from rivers to be used for inefficient training and inference operations which completely burns the planet. This isn't the first time this was reported either so it is a problem [0] [1] still unsolved for years by the deep learning community over finding production-grade efficient methods that doesn't require tons of water and data centers. Don't give me the 'but it is useful' excuses for deep learning when you have no efficient inference methods or alternative training operations that are efficient and reduces 99.999% of energy usage which other problematic industries such as petrol / diesel cars [2], crypto [3], etc have done so with new efficient alternative technologies available today to counter their CO2 that they have contributed. [0] https://www.standard.co.uk/tech/ai-chatgpt-water-usage-environment-study-b1073866.html https://www.standard.co.uk/tech/ai-chatgpt-water-usage-envir... [1] https://gizmodo.com/chatgpt-ai-water-185000-gallons-training-nuclear-1850324249 https://gizmodo.com/chatgpt-ai-water-185000-gallons-training... [2] https://climate.mit.edu/ask-mit/are-electric-vehicles-definitely-better-climate-gas-powered-cars https://climate.mit.edu/ask-mit/are-electric-vehicles-defini... [3] https://consensys.net/blog/press-release/ethereum-blockchain-eliminates-99-99-of-its-carbon-footprint-overnight-after-a-successful-merge-according-to-new-report/ https://consensys.net/blog/press-release/ethereum-blockchain...