i'll believe in orbital data centers as soon as i see one in orbit, until then it is just bullshit

[arrjay@tacobelllabs.net]

@ariadne Orbital Data Center (as relating to the band Orbital)

[ariadne@social.treehouse.systems]

like, i dunno man, i think datacenters weigh a lot.

i don't think we have the rocket to launch something that would actually be meaningful at scale.

starship keeps exploding

[miss_rodent@girlcock.club]

@ariadne I'm... really curious how they expect these to work. A lot of the physicists, engineers, other sciencey types I know are... uh... Skeptical ... about them even being vaguely useful at an actually-achievable scale.

[miss_rodent@girlcock.club]

@ariadne Like there are *so many* fundamental problems to deal with - you need radiation-resistant everything, you need enough bandwidth to get the data in an out, you need cooling - space isn't *cold* it's *empty* and that behaves VERY differently for cooling - how are you going to power them? Solar panels in space is hard to scale. How long do they expect to run for? That radiation problem never goes away... etc. etc.

[brouhaha@mastodon.social]

@ariadne
It's just tulip mania, again.

[ariadne@social.treehouse.systems]

@ireneista i have a more fundamental concern: servers weigh a lot. like the average DC probably has like hundreds of tons in servers alone.

[arrjay@tacobelllabs.net]

@ireneista @ariadne I see none of the concerns about anything resembling a stock server (even a hyperscaler one) not rattling itself apart getting there.

[me@f3d1.thom.cat]

@ariadne yeah but think how much money you'd make as a remote hands tech

[lanodan@queer.hacktivis.me]

@ariadne I'll start believing in orbital data centers when we'll be able to remote manage servers without remote hands in practice.

[ariadne@social.treehouse.systems]

@me nah these hyperscalers just abandon broken servers and let them rot. it's sad.

[joshbressers@infosec.exchange]

@ariadne @ireneista bUt tHeRe iZ nO gRaViTy iN sPaCE!!!!!!!!

[ariadne@social.treehouse.systems]

@joshbressers @ireneista there's gravity here

[rho@social.treehouse.systems]

@ireneista @ariadne if you can power it you can probably cool it, the math works out to roughly the same areal space as the arrays. everything else about orbial dc's is more unrealistic in our opinion

[megmac@social.treehouse.systems]

@ariadne @ireneista I think the idea is to put a lot of smaller units of compute into a huge starlink-style cluster that communicate with each other as well as the ground.

But of course now you're adding extra launch costs, distributed computing *in space*, and many other new problems on top of the cooling one.

[ariadne@social.treehouse.systems]

@megmac @ireneista even if it is not as a single unit, it has to be launched into space

[ireneista@adhd.irenes.space]

@ariadne @megmac you're right, yeah, the tyranny of the rocket equation and all that. there are many serious obstacles and that's one of them; we personally focus on heat dissipation because the answer to "that will cost so much it will bankrupt the world" is "that's okay, this is our destiny! do it anyway!" whereas the answer to "we don't know how" is, at least, some form of "oh okay we'll have to invent something new then"

[ireneista@adhd.irenes.space]

@ariadne @megmac there's this very grim calculus that informs which objection we choose to focus on when trying to make an argument that we want to be taken seriously by people ideologically distant from ourselves

[tknarr@mstdn.social]

@ariadne Let 'em try to put the proposed Stratos data center in Utah into orbit. If they manage it, we'll have no problem with Moonbase Alpha (without the nuclear waste dumps).

[cerement@social.targaryen.house]

@ariadne

farmers across the midwest: It’s raining RAM chips!

[dalias@hachyderm.io]

@ariadne Doesn't matter if you could even. It'll throttle to 10 Mhz because it's impossible to shed any heat.