"SpaceX is hiring to turn Elon Musk's plan for data centers in space a reality"
-
"SpaceX is hiring to turn Elon Musk's plan for data centers in space a reality"
The fact that statements like this don't immediately crash the stock value of his companies proves that the market is both irrational and a terrible way to organize an economy
The only way this could possibly sound like a good idea is you know next to nothing about science
@ekis I can’t help wondering if the real point here is to have data centers that aren’t under the jurisdiction of any particular government, whose owners answer to no one. Plenty of scary outcomes beyond the obvious “AI CSAM in space” punchline…
-
"Sam Altman, CEO of OpenAI, last summer considered buying rocket company Stoke Space to put data centers in orbit, according to the Wall Street Journal"
Seriously, these companies should have lost value when their leaders and spokespeople say something this fucking stupid. Its without question the most embarrassing time to be alive
These people don't have a real education. They just had a lot of money to hire yes men and can't tell when they're being conned on science.
-
"SpaceX is hiring to turn Elon Musk's plan for data centers in space a reality"
The fact that statements like this don't immediately crash the stock value of his companies proves that the market is both irrational and a terrible way to organize an economy
The only way this could possibly sound like a good idea is you know next to nothing about science
As low as $10/kilo...
... Viable yes.Good idea? No...
... Unless you don't want hungry mobs smashing your datacentreThe #broligarchs are working on a total societal collapse
-
As low as $10/kilo...
... Viable yes.Good idea? No...
... Unless you don't want hungry mobs smashing your datacentreThe #broligarchs are working on a total societal collapse
@n_dimension @ekis They intend to put our new AI overlords in orbit where we cannot mob up and smash them.
Elon Musk has rockets and satellites and AI. Now he just needs to buy a nuclear company and he can be a full-on Bond villain. Let's do Moonraker for real, but with crypto instead of gold.
-
@n_dimension @ekis They intend to put our new AI overlords in orbit where we cannot mob up and smash them.
Elon Musk has rockets and satellites and AI. Now he just needs to buy a nuclear company and he can be a full-on Bond villain. Let's do Moonraker for real, but with crypto instead of gold.
-
@n_dimension @ekis So which movie predicts the future then? Forbin Project? Iron Man? Moonraker? i, Robot? Or Terminator 2?
-
As low as $10/kilo...
... Viable yes.Good idea? No...
... Unless you don't want hungry mobs smashing your datacentreThe #broligarchs are working on a total societal collapse
How do you disperse heat? How do you get people up there to maintain it? How do you power the damn thing? How do you get it up there intact? How do you assemble it without an atmosphere? How do you shield against micrometeors? And pieces of other satellites?
It's not viable, they're just scamming people with nonsense. We'll end up with very expensive rocks in orbit, and somehow they still get all our money with no consequences.
CC: @ekis@mastodon.social -
How do you disperse heat? How do you get people up there to maintain it? How do you power the damn thing? How do you get it up there intact? How do you assemble it without an atmosphere? How do you shield against micrometeors? And pieces of other satellites?
It's not viable, they're just scamming people with nonsense. We'll end up with very expensive rocks in orbit, and somehow they still get all our money with no consequences.
CC: @ekis@mastodon.socialAll good questions!
I have had this exact exchange a few days ago (god #tusky non-existant search sucks.
I shall preface this by saying Felon Muscovite is a dick and every time his Nazi rocket blows up, I rejoice.
But a data centre in space is feasible...
Not now, but when the Starship is viable.Here is the response to your questions!
Wulfy—Speaker to the machines (@n_dimension@infosec.exchange)
Attached: 1 image @Bern@aus.social @NewtonMark@eigenmagic.net @cstross@wandering.shop I like folks who do the math! Your eyeball numbers are slightly off my eyeball numbers (you inspired me to check 🙂) The Chinese have been experimenting with Hainan submersible data centres for 3 years now. I believe they have 3 in operation. Using that as an example, 400 servers with inference engine heatload; 1.5 MW = ~30,000m² This translates to two football field sized Louvre arrays. x10+ more than ISS array. However, there are two more effective space cooling systems. One is droplet radiator which would reduce the area to only ~1900m² for the same heat-load, the droplet radiator area to existing ISS dimensions (though diff tech) - This tech is operational Second more effective solution is a nuclear reactor/turbine (demonstrated). The Nuclear Brayton solution only needs ~150m² radiator for the same heatload. Before tRump infested our planet, I would say "good luck with that", but since we now live in times where international laws mean jack shit, getting a nuke up there is an option. BTW: Both solutions are significantly lighter than ISS passive fins. However, I think it's super bad idea because of trash up there, #kesslersyndrome and its just a resource sink for Ketamine boy now that his #occupymars missions are dead and burried. TLDR; Orbital #AI #Datacenter are entirely technologically viable. But, sinking them in the sea is way easier, safer and cheaper. Disclaimer: Back of the envelope calculations. #space
Infosec Exchange (infosec.exchange)
-
"SpaceX is hiring to turn Elon Musk's plan for data centers in space a reality"
The fact that statements like this don't immediately crash the stock value of his companies proves that the market is both irrational and a terrible way to organize an economy
The only way this could possibly sound like a good idea is you know next to nothing about science
True - but remember that what 'investors' in stock markets are looking for is not anything in the real economy - just whether the stock values go up, which is a result only of how many others want to buy.
The gamble is not on the company - it's on other people's psychology.
The stocks in question were sold by the company long ago - these are secondhand trades. It's become so easy to move stock 'investments' that what actually happens in a year's time, let along 5 years, is irrelevant - you'd have made your money and sold by then. (In the end, of course, somebody will get burnt - but who cares, as long as it's not you ?)
-
True - but remember that what 'investors' in stock markets are looking for is not anything in the real economy - just whether the stock values go up, which is a result only of how many others want to buy.
The gamble is not on the company - it's on other people's psychology.
The stocks in question were sold by the company long ago - these are secondhand trades. It's become so easy to move stock 'investments' that what actually happens in a year's time, let along 5 years, is irrelevant - you'd have made your money and sold by then. (In the end, of course, somebody will get burnt - but who cares, as long as it's not you ?)
@GeofCox @ekis Yes. It's FOMO driven by greed: a lot of investors know there's a bubble, they just don't want to miss out on the short-term profits. So long as it's someone else down the line who gets burned (as you put it), they'll carry taking on ever bigger risks, even until...oops, it ended up being them who got burned.
Casino capitalism. A lot of these suits are not much different from the poor sod in Ladbrokes who can't help but put all his winnings straight back down on longer odds.

️ -
"SpaceX is hiring to turn Elon Musk's plan for data centers in space a reality"
The fact that statements like this don't immediately crash the stock value of his companies proves that the market is both irrational and a terrible way to organize an economy
The only way this could possibly sound like a good idea is you know next to nothing about science
Stock markets are casinos.
-
"Sam Altman, CEO of OpenAI, last summer considered buying rocket company Stoke Space to put data centers in orbit, according to the Wall Street Journal"
Seriously, these companies should have lost value when their leaders and spokespeople say something this fucking stupid. Its without question the most embarrassing time to be alive
@ekis He even looks like a village idiot who would crow about putting datacenters in space while being covered in spilled red wine
-
All good questions!
I have had this exact exchange a few days ago (god #tusky non-existant search sucks.
I shall preface this by saying Felon Muscovite is a dick and every time his Nazi rocket blows up, I rejoice.
But a data centre in space is feasible...
Not now, but when the Starship is viable.Here is the response to your questions!
Wulfy—Speaker to the machines (@n_dimension@infosec.exchange)
Attached: 1 image @Bern@aus.social @NewtonMark@eigenmagic.net @cstross@wandering.shop I like folks who do the math! Your eyeball numbers are slightly off my eyeball numbers (you inspired me to check 🙂) The Chinese have been experimenting with Hainan submersible data centres for 3 years now. I believe they have 3 in operation. Using that as an example, 400 servers with inference engine heatload; 1.5 MW = ~30,000m² This translates to two football field sized Louvre arrays. x10+ more than ISS array. However, there are two more effective space cooling systems. One is droplet radiator which would reduce the area to only ~1900m² for the same heat-load, the droplet radiator area to existing ISS dimensions (though diff tech) - This tech is operational Second more effective solution is a nuclear reactor/turbine (demonstrated). The Nuclear Brayton solution only needs ~150m² radiator for the same heatload. Before tRump infested our planet, I would say "good luck with that", but since we now live in times where international laws mean jack shit, getting a nuke up there is an option. BTW: Both solutions are significantly lighter than ISS passive fins. However, I think it's super bad idea because of trash up there, #kesslersyndrome and its just a resource sink for Ketamine boy now that his #occupymars missions are dead and burried. TLDR; Orbital #AI #Datacenter are entirely technologically viable. But, sinking them in the sea is way easier, safer and cheaper. Disclaimer: Back of the envelope calculations. #space
Infosec Exchange (infosec.exchange)
Orbital #AI #Datacenter are entirely technologically viable. But, sinking them in the sea is way easier, safer and cheaper.
So... it's physically and economically inviable, and just a resource sink for Ketamine Boy? I have no particular objections to submersible data centers, other than they're a bad idea, you're a bad person for considering it, and I'll do everything I can to ruin you so you can't keep hurting people like this, but that's true of anyone attempting to build data centers anywhere.
Technologically viable is just another way of saying technically possible but infeasible. Dyson spheres are technologically viable.
CC: @ekis@mastodon.social -
Orbital #AI #Datacenter are entirely technologically viable. But, sinking them in the sea is way easier, safer and cheaper.
So... it's physically and economically inviable, and just a resource sink for Ketamine Boy? I have no particular objections to submersible data centers, other than they're a bad idea, you're a bad person for considering it, and I'll do everything I can to ruin you so you can't keep hurting people like this, but that's true of anyone attempting to build data centers anywhere.
Technologically viable is just another way of saying technically possible but infeasible. Dyson spheres are technologically viable.
CC: @ekis@mastodon.socialNot saying they are economically viable.
But there are two advantages for submersible/space #datacentes
1. They are harder to reach by the coming angry jobless mobs.
2. Both solutions address cooling which is one of the "AI bad for the planet" critique from the #antiai folk
And if Datacentres squeak you, you probably should look too closely what "the cloud" is

-
R relay@relay.infosec.exchange shared this topic