Ancestors

Written by Kate Morley on 2024-12-12 at 11:21

Updated my e-mail signature for the generative AI era:

=> View attached media

=> More informations about this toot | More toots from kate@fosstodon.org

Written by Bas Schouten on 2024-12-12 at 13:20

@kate Actually a ChatGPT query costs between 2 and 10 Wh. If someone were using 15 minutes to respond to your mail on a PC and monitor using ~20W it may very well be better for the environment if they did use LLMs :p. This stuff is pretty hard.

=> More informations about this toot | More toots from Schouten_B@mastodon.social

Written by Petra van Cronenburg on 2024-12-12 at 15:37

@Schouten_B You forgot to include the hidden energy and water costs for training? @kate

=> More informations about this toot | More toots from NatureMC@mastodon.online

Written by Bas Schouten on 2024-12-12 at 17:08

@NatureMC @kate That is amortized when you use it, so not a good argument not to use it. Unless you believe that not using it can make it go away, but that seems extremely unlikely given the scale at which people are already using it and the upward trend in that usage.

=> More informations about this toot | More toots from Schouten_B@mastodon.social

Toot

Written by Bas Schouten on 2024-12-12 at 17:12

@NatureMC @kate (by that logic taking the train is worse than an airplane because the carbon footprint of putting down the tracks is way higher than your flight's is ever going to be)

=> More informations about this toot | More toots from Schouten_B@mastodon.social

Descendants

Written by Petra van Cronenburg on 2024-12-12 at 17:33

@Schouten_B You're welcome to continue making fun of me. I'm still waiting for reliable scientific data to back up your bold claims.

PS: You forgot to count installing the infrastructure for flights ... so much for parallel comparisons.

@kate

=> More informations about this toot | More toots from NatureMC@mastodon.online

Written by Bas Schouten on 2024-12-12 at 19:52

@NatureMC @kate I'm not making fun of anyone. I'm simply pointing out the flaw in the argument. LLMs are much more efficient than humans at many task. If a single GPT4 query saves you 20 minutes behind an average 27 inch monitor, that's an obvious net positive for the environment.

The installation for the infrastructure of flights has a vastly lower footprint than for trains. Luckily trains are a lot more efficient than planes of course. Which is why using a train is good (for <1000km).

=> More informations about this toot | More toots from Schouten_B@mastodon.social

Written by Petra van Cronenburg on 2024-12-13 at 10:39

@Schouten_B I'll answer at your other points. Here a short explanation for my question (it was not an argument): I asked by curiosity. Because we forget the hidden costs of products much too often. If in Europe, I buy for 10 E strawberries in decembre or for 10 E winter apples from the farmer in my neighbourhood, it's 10 E = 10 E. But the hidden costs are quite different. I don't pay them directly but our society or people in other countries pay for it. The CO2 imprint is different, too.

@kate

=> More informations about this toot | More toots from NatureMC@mastodon.online