The first statement is not even wholly true. While training does take more, executing the model (called “inference”) takes much, much more power than non-AI search algorithms, or really any traditional computational algorithm besides bogosort.
Big Tech weren’t doing the best they possibly could transitioning to green energy, but they were making substantial progress before LLMs exploded on the scene because the value proposition was there: traditional algorithms were efficient enough that the PR gain from doing the green energy transition offset the cost.
Now Big Tech have for some reason decided that LLMs represent the biggest game of gambling ever. The first to find the breakthrough to AGI will win it all and completely take over all IT markets, so they need to consume as much as they can get away with to maximize the probability that that breakthrough happens by their engineers.
Yeah, I ran some image generators on my RTX2070, and it took a solid minute at full power to do it. Sure, it’s not a crazy amount, but it’s not like it’s running on your iPhone.
Nothing more than playing a game for that minute
there aren’t many games that tax your gpu like inference, where it’s pegged for the entire time. i have a power usage tracker on my desktop because my gpu is stupidly powerhungry, and it uses way more power doing inference than playing a modern graphics-intensive game.
you can get a very small generator running on a modern phone if you want a grainy 400x400 piece of anime trash
You have more brians than any AI shill will ever be able to emulate.
brians
You rang?
Only takes two brians to summon you?
Convenient!
Thirty Brians agree…
Cue badass bass intro song
Uh uh…
Brian’s World! Brians World!
Radical!
I don’t use autocorrect, because misspellings and misstypes are usually more legible than what ever ‘other’ word autocorrect picks.
… though I’m leaving this one in. A gaggle of Brians is better than any AI shit.
God autocorrect is so bad on my phone
I had it turned off for the longest time and it got reset recently back to on and I’m giving it a few days to see if anything has improved.
It sure likes changing words it doesn’t have saved to other words instead of just doing an underline or something helpful.
Yeeep.
Autocorrect will out whole ass entirely wrong words in and significantly change the conveyed meaning, often into gibberish.
Maenwhlie, yuo cn proably. rea d this quirt ealisly.
Ah, Muphry’s law at work
AI training is less energy than AAA game development of which dozens come out each year.
Something that sucks is slightly less bad for the environment than something that also sucks. Wow very good argument
^ whataboutism
Is it wrong to point out hypocrisy? Is also it wrong to mock SUV owners who complain that buses burn gasoline/diesel?
It’s that something you have a source on, or do you just have a feeling about it?
Gpt 4 had a training cost of $78 million. Gta5 cost $300 million. 4000 developers each with the latest GPU burning hundreds of watts per employee to create the assets. A rough estimate of 750watt pc, 4,000 developers, 8 hour a day, 300 days a year, 5 years = 36 giga watt-hours.
https://en.m.wikipedia.org/wiki/GPT-4
https://gamerant.com/rockstar-games-with-longest-development-cycles-ranked/ https://www.tweaktown.com/news/88292/over-6-000-people-worked-on-grand-theft-auto/index.html https://www.sportskeeda.com/gta/how-long-take-make-gta-5-a-brief-history-game-s-developmental-cycle
Off topic: while trying to find my post where I did the math I discovered that Blue Morpho is the name of an AI company. God damn it. AI ruined my Venture Bros reference.
Ignoring the shitty quick maths. Those are energy costs of employing people. Those programmers and artists won’t stop needing AC and a computer if you get rid of the videogame industry, they’ll move to another industry with AC and computers.
Yeah, but Game development does use more energy than a typical office worker. However even treating it as a regular office worker, the training costs are equivalent to 3k office workers but the results of the training are used by far more than 3k. So the energy use of the training is divided by millions of users.
So you’re saying that GTA VI is only going to be used by the developers?
And how does tech companies putting AI results on every interaction of mine count as a user? I never read their bullshit, yet it’s all over my screen, wasting both insane amounts of energy and valuable screen space.
deleted by creator
Except their GPUs wouldn’t be at 100% power all the time. Most of that time will be with programming, texture artistry, planning mechanics etc. I have a feeling that “back of the napkin” calculation is incredibly inaccurate.
100% power all the time
A gaming PC uses more than 750 watts. Plus there’s AC cooling for that which doubles the total power usage. So 750 watts includes not using the CPU all the time. Then there’s the testers on the development team which greatly outnumber the coders. Testers are using the GPU the entire time they’re working.
Lastly the cost comparison of the Rockstar developing GTA5 vs OpenAI training ChatGPT includes all the energy costs because they had to pay for everything. OpenAI is 2k developers compared to Rockstar 4k so it’s not just salaries that make up the difference.
No, a gaming pc ‘can’ use 750 watts. But it won’t be using 750 watts 100% of the time.
Power supplies only draw what they need, not their full rated capacity.GPU’s also downclock and lower their power draw when they aren’t being stressed to their limits.
Lunch breaks cut 30-60 minutes out of those 8 hours.
And that article estimated 5000 employees internationally, which includes management, marketing, programmers, visual artists, musicians, sound engineers, voice actors, qa, translators, external consultants/contractors, and many more.
These are all people who have helped make the game at some point, and does not mean these people were in-office or employed for the entire development process.Very few of those roles require top-end machines, and even fewer require their systems to be running at full speed 100% of the time.
Even if we ignore that that gaming companies have a larger energy footprint, it’s still less energy. Let’s assume its a regular corporation with only accountants and laptops:
6k employees 20 kwatt hours per square foot https://esource.bizenergyadvisor.com/article/large-offices 150 square feet per employee https://unspot.com/blog/how-much-office-space-do-we-need-per-employee/#%3A~%3Atext=The+needed+workspace+may+vary+in+accordance
18,000,000,000 watt hours
vs
10,000,000,000 watt hours for ChatGPT training
https://www.washington.edu/news/2023/07/27/how-much-energy-does-chatgpt-use/
sure, while we’re splitting hairs let’s ban all computers worldwide because they are the cause of the huge increase of energy needs.
Neo-neolithic age, here we come.
At least human beings get paid to do that and presumably at least some of the people playing the games enjoy the experience.
The concept of a video game also isn’t fundamentally built on consumer deception, either.
Good point, we should probably stop doing that. Reject modernity, return to ASCII
