The true cost of GPT-3
Water Consumption of GPT-3 training
Water Consumption of GPT-3 training
Starting the conversation about the environmental impact of the technology we take for granted is a step in the right direction. Most people do not think about the effects of the technology they use. A large majority of computation, from our messaging, our streaming, and now AI assistants, happens at a server out of sight of the user. This disconnect means as a society, we have become detached from the real impact of our technology. With a rapidly worsening climate situation, we need to get the conversation going on ethical and sustainable technology seriously. The data presented here is just a tiny snapshot. This site attempts to get that conversation going, and I hope it inspires others to do the same.
All visualisations were done using Matplot, Pandas and seaborn in Jupyter by Luna Lamb
[1]https://searchengineland.com/calculating-the-carbon-footprint-of-a-google-search-16105#:~:text=In%20terms%20of%20greenhouse%20gases,about%200.2%20grams%20of%20CO2.
[2]https://arxiv.org/ftp/arxiv/papers/2303/2303.06219.pdf
[3]https://www.waterwise.org.uk/save-water/#:~:text=The%20average%20showerhead%20uses%2012,don't%20compromise%20on%20pressure.
[4]https://www.thehindubusinessline.com/blink/know/it-takes-7600-litres-of-water-to-make-your-jeans/article30871977.ece#:~:text=Choose%20what%20you%20like%2C%20but,make%20it%20through%20production%20line
[5]https://www.energysavingtrust.org.uk/sites/default/files/reports/AtHomewithWater%287%29.pdf
[6]https://puiij.com/index.php/research/article/view/39/23
