It doesn’t take that much energy and power to run an LLM or an image generator, and sure, it would take a lot with so many users connecting, across so many servers… but there’s just no way they’re not mining bitcoin. My math might not be mathing but it seems like AI doesn’t justify the power use and it seems like everybody’s lying.
Someone who knows more about this inform me of your opinions
Bitcoin mining doesn’t normally use GPUs. They use dedicated ASICs. Far more effective, and cheaper in every way.
In addition, AI is in the “growth at any cost” phase. There is a TON of investor money to burn, with little need to show future profitability.
If that’s the case, why don’t they pull the Renaissance artist scam and just take a bunch of money and swear they’re working on stuff? If they don’t have to prove anything, why use any energy at all?
Little need, but not no need. They need to have a vague path, and something to show for it.
A vague path and something to show for it means using a bunch of energy for something they’re not even really doing…?
always ask:
- How many people would have to be in on it?
- How likely is it for everybody to keep the secret?
And also ask: do they gain anything from lying about it?
I mean… yeah. It’s always beneficial for people not to know how much money one has… I can’t really imagine a situation where it wouldn’t be.
The big power use is in building/training models, not running them. Not the running them is insignificant either.
Running LLMs do take a lot of power. Trying to run a very small model on my old laptop makes the fans go crazy and power draw must be high given that it gets very hot. Then when you get models with billions upon billions of parameters, you can see how it all adds up. Then you multiply these humongous models with millions of users, as well as consider all the energy required to cool all those servers (fans, pumps, etc).
The biggest power draw in the AI industry is probably training the models, given that it’s quite literally going through the entire internet for every crumb of data and mixing it all together into coherent text and images. Then you see how power grids can be taken down by these power hungry data centers.
If you have seen any videos of people visiting these data centers, you can hear the incredibly loud fans that make it difficult to hear the presenter speaking.
Thats true of all datacenters though, not just AI. 1RU fans are just crazy noisy.
Yeah it’s really noisy and hot when you mine bitcoin and distract everyone by reading their emails, great distraction actually, kudos
Start with he premise that energy is not created or destroyed, it only changes state. Pushing processors to perform a task uses energy and creates heat to rearrange numbers and pixels. Combing through huge numbers of data vits takes a lot of energy. If you’ve ever worked with systems larger than a desktop you know how much they can draw.
My whole home lab uses about 700 watts for basic server duties. To power a single 3090 GPU to run some larger models my PC needed a 1000 watt PSU. Extrapolate that out to a commercial size org and you can see where the major power draw comes in.
That makes sense.
They are both a waste of power. Just because one is less of a waste doesn’t make it a good choice.
I didn’t say it was a good choice, not sure I made any kind of statement you could nitpick or impose judgement upon me for, and if I did, I edited it out because I predicted you being insufferable. “I think they’re mining bitcoin” is not the exact same statement as “bitcoin is good” or “AI is good.” I don’t actually mind either bitcoin or AI, I think like alcohol they’re fine in moderation, but that’s not what my post is about.
