Is there anyway to make it use less at it gets more advanced or will there be huge power plants just dedicated to AI all over the world soon?
will there be huge power plants just dedicated to AI all over the world soon?
Construction has started(or will soon) to convert a retired coal power plant in Pennsylvania to gas power, specifically for data-centers. Upon completion in 2027 it will likely be the third most powerful plant in the US.
The largest coal plant in North Dakota was considering shutting down in 2022 over financial issues, but is now approved to power a new data-center park.
Location has been laid out for a new power plant in Texas, from a single AI company you’ve probably never heard of.
And on it goes.
Data is the new oil. Collecting it, refining it, and distributing it.
will there be huge power plants just dedicated to AI all over the world soon?
It takes time to build a power plant. A more realistic scenario is that we’ll continue as we have: AI centers will be built wherever local governments approve them for the taxes, without regard for the strain they put on the aging electrical grid and, given the massive amount of electricity they need, everyone’s electrical bill will just massively increase.
They’ve been building a large number of data centers and AI centers in Virginia, and it’s been straining and raising prices across the entire PJM interconnector region, to the point where at least a couple states are considering leaving it. Microsoft has bought the rights to and is reactivating part of the Three Mile Island nuclear plant, because they wanted dedicated power, and they’re still going to be pulling power from the grid.
Also water, they consume heaps of fresh water which is used for important meat bag things like, oh I don’t know, eating and drinking perhaps.
No one is really challenging them on this, but water scarcity is going to be a big deal as climate change worsens.
Cook the planet and take all the water.
Isn’t water mainly used for cooling? I think you can still drink that water, unless they pump it full of chemicals.
the current cooling paradigm is to basically spray mist into the air inlets of a data center to make the air able to carry more heat. the hot, moist air is then vented to atmosphere. so the water is lost until it rains again.
So… the water is never lost.
Water is never lost.
The problem is getting, filtering, and purifying the water so it can be used again.
sure, just like how water is not lost when you take a piss in the woods. it’s just not reusable without significant energy expenditure.
It’s mostly the training/machine learning that is power hungry.
AI is essentially a giant equation that is generated via machine learning. You give it a prompt with an expected answer, it gets run through the equation, and you get an output. That output gets an error score based on how far it is from the expected answer. The variables of the equation are then modified so that the prompt will lead to a better output (one with a lower error).
The issue is that current AI models have billions of variables and will be trained on billions of prompts. Each variable will be tuned based on each prompt. That’s billions to the power of billions of calculations. It takes a while. AI researchers are of course looking for ways to speed up this process, but so far it’s mostly come down to dividing up these billions of calculations over millions of computers. Powering millions of computers is where the energy costs come from.
Unless AI models can be trained in a way that doesn’t require running a billion squared calculations, they’re only going to get more power hungry.
This is a pretty great explanation/simplification.
I’ll add that because the calculations rely on floating point math in many cases, graphics chips do most of the heavy processing since they were already designed for this pipeline in mind with video games.
That means there’s a lot of power hungry graphics chips running in these data centers. It’s also why NVidia stock is so insane.
OpenAI noticed that Generative Pre-trained Transformers get better when you make them bigger. GPT-1 had 120 million parameters. GPT-2 bumped it up to 1.5 billion. GPT-3 grew to 175 billion. Now we have models with over 300 billion.
To run, every generated word requires doing math with every parameter, which nowadays is a massive amount of work, running on the most power hungry top of the line chips.
There are efforts to make smaller models that are still effective, but we are still in the range of 7-30 billion to get anything useful out of them.
My understanding is that traditional AI essentially takes a bruteforce approach to learning and because it is hardwired, its ability to learn and make logical connections is impaired.
Newer technologies like organic computers using neurons can change and adapt as it learns, forming new pathways for information to travel along, which reduces processing requirements and in turn, reduces power requirements.
Machine learning always felt like a very wasteful way to utilize data. Even with ridiculous quantities of it, and the results are still kinda meh. So just dump in even more data, and you get something that can work.
Lower voltage chip advancement along with better cooling options may come along some day.
They should consider building their super centers underwater in places like Iceland.
Thats only a short term solution, global warming will negate those benefits.
Southern ocean currents just reversed and will likely cause rapid warming of water temps.
Southern Ocean circulation reversed
Southern Ocean current reverses for first time, signalling risk of climate system collapse
France and Switzerland just had to shutdown their nuclear reactors due to the water sources they use for cooling being too warm.
France and Switzerland shut down nuclear power plants amid scorching heatwave
None of that is terrifying at all /s.
imagine that to type one letter, you need to manually read all unicode code points several thousand times. When you’re done, you select one letter to type.
Then you start rereading all unicode code points again for thousands of times again, for the next letter.
That’s how llms work. When they say 175 billion parameters, it means at least that many calculations per token it generates
That’s how llms work. When they say 175 billion parameters, it means at least that many calculations per token it generates
I don’t get it, how is it possible that so many people all over the world use this concurrently, doing all kinds of lengthy chats, problem solving, codegeneration, image generation and so on?
that’s why they need huge datacenters and thousands of GPUs. And, pretty soon, dedicated power plants. It is insane just how wasteful this all is.
So do they load all those matrices (totalling to 175b params in this case) to available GPUs for every token of every user?
yep. you could of course swap weights in and out, but that would slow things down to a crawl. So they get lots of vram (edit: for example, an H100 has 80gb of vram)
I also asked ChatGPT itself, and it listed a number of approaches, and one that sounded good to me is to pin layers to GPUs, for example we have 500 GPUs: cards 1-100 have permanently loaded layers 1-30 of AI, cards 101-200 have permanently loaded layers 31-60 and so on, this way no need to frequently load huge matrices itself as they stay in GPUs permanently, just basically pipeline user prompt through appropriate sequence of GPUs.
Supercomputers once required large power plants to operate, and now we carry around computing devices in out pockets that are more powerful than those supercomputers.
There’s plenty of room to further shrink the computers, simplify the training sets, formalize and optimize the training algorithms, and add optimized layers to the AI compute systems and the I/O systems.
But at the end of the day, you can either simplify or throw lots of energy at a system when training.
Just look at how much time and energy goes into training a child… and it’s using a training system that’s been optimized over hundreds of thousands of years (and is still being tweaked).
AI as we see it today (as far as generative AI goes) is much simpler, just setting up and executing probability sieves with a fancy instruction parser to feed it its inputs. But it is using hardware that’s barely optimized at all for the task, and the task is far from the least optimal way to process data to determine an output.
Supercomputers once required large power plants to operate, and now we carry around computing devices in out pockets that are more powerful than those supercomputers.
This is false. Supercomputers never required large [dedicated] power plants to operate.
Yes they used a lot of power, yes that has reduced significantly, but it’s not at the same magnitude as AI
It is also a very large data set it has to go through the average English speaker knows 40kish words and it has to pull from a large data set and attempt to predict what’s the most likely word to come next and do that a hundred or so times per response. Then most people want the result in a very short period of time and with very high accuracy (smaller tolerances on the convergence and divergence criteria) so sure there is some hardware optimization that can be done but it will always be at least somewhat taxing.
This is an astute answer. Bravo.
