Kratzkopf

joined 2 years ago
[–] Kratzkopf@discuss.tchncs.de 28 points 21 hours ago (3 children)

Well, that's progress, innit? After you read A and B you set out to improve things further and it worked. That's why you publish it.

(But don't get me started on systematic problems in academic publishing which stop people from publishing their helpful results about not succeeding and also exaggerating the importance of their findings)

[–] Kratzkopf@discuss.tchncs.de 0 points 2 days ago (1 children)

Which vector isn't pointing tho?

[–] Kratzkopf@discuss.tchncs.de 1 points 3 days ago

I do not think that is likely because as far as I know (please correct me if I'm wrong) these image generations are mostly done on servers which use the cheapest energy mix. But even if they use a 'green' server or run stable diffusion locally and power it by their rooftop solar, the energy for this picture would otherwise have gone to the grid and replaced some fossil energy source. One AI image takes approximately 3Wh of energy (enough to boil ~86 cups of tea). With a gas power plant this produces about 6mg of CO2. That doesn't sound like a lot but it adds up. We really need to reduce emissons wherever we can and this would be an easy place to do so.

[–] Kratzkopf@discuss.tchncs.de 2 points 5 days ago (2 children)

It would have been a funny joke without the picture, too. For me the unnecessary energy consumption of making a picture by AI is the more relevant criticism here.