noticing it more and more. i like to play retro video games but i don’t actually do it that often. 5 years or so ago i got a SNES mini and a playstation classic and one of those anbernac handhelds in the span of a year or two, so I was spending a lot of time reading lists online of games people thought were notable, underrated, good games for genre newbies, etc.
i have been playing games again recently so i have again been looking up games and the difference in content you get now is astounding. five years ago if you searched something like “best nes RPGs” or “obscure ps1 games” you would find lovingly handcrafted lists and articles by people who were passionate about it and wanted to share, make readers laugh, or ignite interest in something. Now there’s like 20 different sites that each have ai generated “best (genre) games for (system)” lists for every system and genre combination possible, with generic game descriptions, list orders likely cribbed from one of those ranking sites, and nonsensical filler copy (“every RPG enthusiast loves the N64” type words just mashed together)
photographs are also no fun to take or look at anymore, accelerated by new ai image generation but honestly ever since smartphone cameras started automatically editing the shot out of your picture before it even showed it to you.
when i was a kid i wanted to be an author, glad i just got depressed and useless and never pursued it, considering what that space looks like today.
internet was a mistake
https://www.nature.com/articles/d41586-024-00478-x
https://www.technologyreview.com/2023/12/05/1084417/ais-carbon-footprint-is-bigger-than-you-think/
Doesn’t seem that marginal to me, especially when those margins still affect the overall outcome and accelerate the worsening of environmental conditions.
With the exception of the cited pre-print on water usage, those article don’t seem to quantify the actual use.
The water usage is high from including scope-2 water withdrawal and consumption, which is the water usage of power plants used to generate electricity and not primary data centre cooling. Looking only at primary cooling, Google used 25 gigalitres in 2022 - for comparison, Arizona uses 8,600 gigalitres of water a year.
A couple percent is marginal - if the average temperature increase is 4.3 degrees C or 4.28 degrees C in 2100, it doesn’t really impact the outcome.
I think the more substantive impact is the billions in resources (financial, material and labour) that are being dumped into yet another bazinga scheme instead of things we know work and will help slow climate change.
How’s this? Keep in mind the corpos pushing this shit are trying to keep the actual costs to themselves.
https://www.vox.com/climate/2024/3/28/24111721/climate-ai-tech-energy-demand-rising
It could become less marginal in the future, but I don’t know how much juice is left to be squeezed out of AI hype.
Oh well, another $100 billion into GPUs and data centres and maybe they’ll find something else to fuel the hype train for a bit longer.
Microsoft seems to have some high high hype hopes:
https://techcrunch.com/2024/09/20/microsoft-taps-three-mile-island-nuclear-plant-to-power-ai/
And of course King Bazinga keeps trying to expand his private stake in his mirror, mirror, on the wall:
https://www.reuters.com/business/environment/musks-xai-operating-gas-turbines-without-permits-data-center-environmental-group-2024-08-28/