The citation for the "surprisingly large carbon footprint" [0] is crazy. The paper alleges that a car, including fuel, has a lifetime footprint of 126,000 lbs of CO2e, and training a big NN transformer consumes 626,155 lbs of CO2e, almost a 5-fold increase.
The carbon footprint idea seems flawed to me. It always just depends how far you look. There are machines making machines making machines and so on, the chain is really long and those early machines in the chain, which may have been necessary to create given product, may not even exist anymore. They might have existed before the product was even invented.
Note that the whole idea was mostly popularized by BP [1][2].
If only the carbon footprint of a human was the body carbon.
Modern humans have a very heavy carbon footprint, especially in the US. Think of all the things you do and consume and all the carbon involved all thorough the chain. It's a big number. Computers are extremely efficient compared to that.
Carbon footprint depends on energy sources, not population density. 100,000 humans supplied with coal-fired electricity have a radically different carbon footprint from the same populations supplied with solar PV and wind electricity, and it doesn't matter whether they live in isolated households or in one skyscraper.
It goes even beyond this and comments like yours are important to stress that there is a lot of nuance. If you count carbon emissions related to building the plant, the enrichment facility, the storage facility, the mining, the lonf-term maintenance, the decomissioning, etc you end with quite a carbon footprint as (so far) that all relies on diesel.
reply