Bitcoin needs ASICs to run it. GPUs haven’t been viable there for a long time, and ASICs aren’t useful for anything other than the thing they’re designed to do.
Ethereum used GPUs until it went proof-of-stake, but it was always smaller than Bitcoin.
Nothing else is big enough to have caused a bubble.
Most of the AI training is being done in brand new datacenters on brand new GPUs. Those Ethereum GPUs mostly got dumped on eBay.
ASICs are used to mine crypto, but the ledger itself is on a regular server, more or less. It’s also true that LLMs are trained on High throughput ASICs. I don’t know what that volume looks like rn wrt AI, but I imagine its still in high demand.
The number of actual active coins has barely changed since the peak. Its a little higher, and there are more coins now, but they are significantly less active. Active trading on exchanges during the peak was about 25T, now its around 18T, a decrease of over 1/4. Are you arguing that 7 trillion dollars in annual exchange volume is inconsequential? At the peak there were about 400 exchanges, now there’s a little more than 200 active exchanges. Do exchanges not use any computing power? Are people trading bitcoin by hand somewhere? I’d love to see that.
There’s still a good amount of marketing and advertising but it has cooled a great deal. Lots of that business has consolidated and sees significantly less traffic = less computing power.
Added to the post covid lockdown decline in demand for computing services, which affected all of the same subjects of concern, there was a fuckload of unused computing power, just sitting around. This computing power represents a ton of investment in infrastructure, and long term plans that could not be stopped. AI is the justification to buy up a bunch of cheap server space available at the time, and continue growth of data centers and chip manufacturing plants (now stalled for 5 years or more). Why are all these companies who are heavily invested in vast server farms and chip manufacturing going so hard for a service that barely works?
The answer isn’t technical, its economic and political. I have no doubt you know more about chips and computing than I do, but if you want to understand what is happening, you have to look at economics and politics. The technical world of these tiny little areas of concerns and continuously manicured definitions is not how you tell what is going on in the world of chips and computing.
Trump is buying 10% stake in Intel, whose stalled manufacturing plant in Ohio was once visited by Biden and heralded at his state of the union. And the justification for it is National Security concerns. The great tech giants all sat right behind Trump at his inauguration. Did you not get the message? Chips and servers are a national security issue, it is incredibly political and as such greatly influences, and is greatly influenced by, an economic system that has to constantly grow greater returns of profit.
Anyway coin maintenance, ancillary services and subsequent economic activity surrounding crypto uses the same kind of computing power as AI. Its not like once a chip is mined with a ASIC GPU it stops using computing power, quite the opposite, but it does stop using ASIC computing power.
The bubble isn’t just the thing itself, its all the services and infrastructure that built up around it. Crypto doesn’t have to be awful, its just a tool, a ledger. But it exists in its current form because of what it creates, what kind of economic activity it stimulates. My point is that all the grifty ass economic hype activity around crypto just moved to AI
But I do appreciate the clarification, it gives me a chance to refine my perspective
The crypto bubble is the ai bubble. AI was the answer to “what are we going to do with all these chips and servers now that crypto crashed?”
Bitcoin needs ASICs to run it. GPUs haven’t been viable there for a long time, and ASICs aren’t useful for anything other than the thing they’re designed to do.
Ethereum used GPUs until it went proof-of-stake, but it was always smaller than Bitcoin.
Nothing else is big enough to have caused a bubble.
Most of the AI training is being done in brand new datacenters on brand new GPUs. Those Ethereum GPUs mostly got dumped on eBay.
ASICs are used to mine crypto, but the ledger itself is on a regular server, more or less. It’s also true that LLMs are trained on High throughput ASICs. I don’t know what that volume looks like rn wrt AI, but I imagine its still in high demand.
The number of actual active coins has barely changed since the peak. Its a little higher, and there are more coins now, but they are significantly less active. Active trading on exchanges during the peak was about 25T, now its around 18T, a decrease of over 1/4. Are you arguing that 7 trillion dollars in annual exchange volume is inconsequential? At the peak there were about 400 exchanges, now there’s a little more than 200 active exchanges. Do exchanges not use any computing power? Are people trading bitcoin by hand somewhere? I’d love to see that.
There’s still a good amount of marketing and advertising but it has cooled a great deal. Lots of that business has consolidated and sees significantly less traffic = less computing power.
Added to the post covid lockdown decline in demand for computing services, which affected all of the same subjects of concern, there was a fuckload of unused computing power, just sitting around. This computing power represents a ton of investment in infrastructure, and long term plans that could not be stopped. AI is the justification to buy up a bunch of cheap server space available at the time, and continue growth of data centers and chip manufacturing plants (now stalled for 5 years or more). Why are all these companies who are heavily invested in vast server farms and chip manufacturing going so hard for a service that barely works?
The answer isn’t technical, its economic and political. I have no doubt you know more about chips and computing than I do, but if you want to understand what is happening, you have to look at economics and politics. The technical world of these tiny little areas of concerns and continuously manicured definitions is not how you tell what is going on in the world of chips and computing.
Trump is buying 10% stake in Intel, whose stalled manufacturing plant in Ohio was once visited by Biden and heralded at his state of the union. And the justification for it is National Security concerns. The great tech giants all sat right behind Trump at his inauguration. Did you not get the message? Chips and servers are a national security issue, it is incredibly political and as such greatly influences, and is greatly influenced by, an economic system that has to constantly grow greater returns of profit.
Anyway coin maintenance, ancillary services and subsequent economic activity surrounding crypto uses the same kind of computing power as AI. Its not like once a chip is mined with a ASIC GPU it stops using computing power, quite the opposite, but it does stop using ASIC computing power.
Actually…
Crypto mining requires high compute throughput.
LLM’s require high VRAM.
Very different objectives and hardware in mind.
The bubble isn’t just the thing itself, its all the services and infrastructure that built up around it. Crypto doesn’t have to be awful, its just a tool, a ledger. But it exists in its current form because of what it creates, what kind of economic activity it stimulates. My point is that all the grifty ass economic hype activity around crypto just moved to AI
But I do appreciate the clarification, it gives me a chance to refine my perspective
Running the model? Sure. Training the model still requires high compute throughput.