Mastodawn

MistaPlanet Feb 10

Datacenters in space are a terrible, horrible, no good idea.

“I am a former NASA engineer/scientist with a PhD in space electronics. I also worked at Google for 10 years, in various parts of the company including YouTube and the bit of Cloud responsible for deploying AI capacity, so I'm quite well placed to have an opinion here.
The short version: this is an absolutely terrible idea, and really makes zero sense whatsoever.”

Have a look, @sundogplanets

https://taranis.ie/datacenters-in-space-are-a-terrible-horrible-no-good-idea/

Datacenters in space are a terrible, horrible, no good idea.

There is a rush for AI companies to team up with space launch/satellite companies to build datacenters in space. TL;DR: It's not going to work.

Taranis

Richard Nairn Feb 10

@jodmentum Yes, I can’t imagine how the economics could make sense. One proposed data center for Alberta is looking at 1Gw of power and two million sq feet. How could you construct anything comparable. Heat dissipation component failure and upgrades, bandwidth. What are the advantages of space…

0xC0DEC0DE07EA Feb 10

@RichardNairn @jodmentum Wall-E taught me that “There’s plenty of space out in space.”

Crucifixion's a Doddle Feb 10

@RichardNairn @jodmentum I suspect part of the attraction for these lunatics is that space is “out of the reach of regulation”.

Laying aside that it literally isn’t… why d’you reckon they find that idea attractive?

Skjeggtroll Feb 10

@drunkenmadman @RichardNairn @jodmentum

It's not, though. Sure, you can't physically raid a satellite in orbit, but you _can_ raid the ground station and subpoena any encryption keys required. Satellites in space are no more outside of national jurisdiction than ships on the high seas are. Musk has a history of ever increasingly harebrained schemes -- any pie-in-the-sky dream he can sell to investors with the promise of unccountable riches at the ens of the rainbow. It's one of those.

@drunkenmadman @RichardNairn @jodmentum it all makes me imagine some dystopian movie where they put data centers in space to house AI that controls drones that control the population. So the population can’t get sick of it and burn the data centers.

Ziege Lightbocks Feb 10

Could start with Lords of the Middle Dark by #JackChalker, but I'm sure there are others.

@drunkenmadman @RichardNairn @jodmentum

evancull Feb 10

@shadows you're close, think more along the lines of low orbit satellite ICBM and Hypersonic interceptors trained on the iron dome and you'd be right on the money, or should I say taxpayers money, forever.

@jodmentum @sundogplanets the worst of tech ideas...

Toni Aittoniemi Feb 10

@jodmentum @sundogplanets Data Centers in space are literally a grift invented for one sole purpose: Push up the stock of Elon Musk’s IPO for a combined SpaceX, XAI, Tesla.

It demonstrates of the terrible memetic power a single centibillionaire can wield.

Our society is not well.

☮ ♥ ♬ 🧑‍💻Feb 10

@jodmentum “A huge, ISS-sized, array could therefore power roughly 200 GPUs”

I’m surprised nobody is sealing GPUs underwater in the ocean. cc
@sundogplanets

Ninad Pundalik Feb 10

@peterrenshaw @jodmentum @sundogplanets have you already read about https://en.wikipedia.org/wiki/Project_Natick?

Project Natick - Wikipedia

Ziege Lightbocks Feb 10

Maybe environmental vacuum cooling is as fiddly and disaster prone as suberged liquid cooling.

☮ ♥ ♬ 🧑‍💻Feb 10

@lightbocks yes, I’m throwing the idea around in my head. Maybe using coastal areas with access to cold weather and seas could be useful? Another idea is looking at passive cooling via buildings?

Ziege Lightbocks Feb 10

Right direction.

Don't forget to add federation, or it's still just silicon colonization doing usual amounts of damage to neighborhoods and towns.

Energy intake and heat dispursion benefit greatly from federation.

Ie, home-manage, community supported solar getting panels or gardens on every roof v.s. destroying gardens and ecosystems for panel farms.

So I'd rather support small servers and help people find the most efficient placement for heat.

Ziege Lightbocks Feb 10

@peterrenshaw
Considering we can and do crowd source protien folding models trying to cure cancers... Big data doesn't really have anything I want, I guess.

Ziege Lightbocks Feb 10

@peterrenshaw
to be clear, when I made my origional comment I was dying inside of @Markiplier going through the beggining, middle, end, and eight stages of greif with Mirabilite/ Decahydrated sodium sulfate mineral/ GlauberSalt.
(pop culture sillyness, spare w.c. labs)

Typical, there's a lot more sand in the math than a rug-seller tells you.
Rug-makers, on the other hand? Half of them won't shut up about dirt -- best to clean rugs though.

this has been a metaphore about any crafting workflow.

DragonBard Feb 10

@jodmentum @sundogplanets @peterrenshaw Microsoft did. The problem was maintenance. The expense of pulling up the pressure vessel to replace failed components was too much.

@jodmentum @sundogplanets What would you consider as a fall back system in very extreme situations like a nuclear war? Is then a bunker system better then in space? Does it even make no sense in rare special cases?

Sure, IF we could lift that much equipment to geosynchronous orbit, and IF we could construct large orbital infrastructure 22,000 km up, and IF we could get fast, effective, stable, high-speed, low-lag communications from ground to orbit, and IF we could maintain that infrastructure daily, and IF we could manage the heat dissipation problem, and ... it might make sense.

Or once we are able to permanently live off-planet.

There are too many ifs for our current level of space tech.

Not to mention: in a nuclear war, ground infrastructure will be destroyed, and any space-based infrastructure would be sitting duck for satellite killers (a ballbearing with attitude).

One day, if we can get our shite together as a planet, but not while big country A has designs on big country B for resource or ideological "needs".

@jodmentum @sundogplanets

@PeterLG @jodmentum @sundogplanets hmm I dont speak about a big facility, but on the moon a small one for the basic things? I realy talk about only needed fallback systems.

There's no such thing as a small data centre, and the moon is a lot further away than Geosynchronous Earth Orbit.

There's also the problem of access: the moon isn't visible in the sky over one part of the Earth (such as the USA) all of the time, so that means multiple ground stations, which means connected infrastructure — which is vulnerable.

Another consideration: after our nuclear war bumps us back a few thousand years, how many millennia will it take for us to get access to that data again? And will it even be understandable to future humankind?

All of these considerations are reasons to become a multiple-planet species sooner rather than later. Stuck on one planet (that we really don't seem to appreciate) means one idiot with the nuclear codes could undo everything. Depressing.

@jodmentum @sundogplanets

@PeterLG @jodmentum @sundogplanets ok then everything that does not fit into the ISS makes no sense. OK, good argument. More than a small Homeserver is then not possible. I maybe confused the word Datacenter. A real center, you are correct. A small Home Server on the ISS, or things like ESHail2, but I think is possible. https://en.wikipedia.org/wiki/Es%27hail_2

Es'hail 2 - Wikipedia

@PeterLG @jodmentum @sundogplanets

By the way...
The dream to leave our solarsystem to live on another planet also is a missleading ilusion. The only way is by frozen gene material and AI Robots who clone and raise the brought species to the new planet. Alot of people think we ever would make it to leave our solar system, wich is not possible. We only can leave it dead or as genetic material only. An Idea of a blackhole as way to reach some speads is an ilusion.

@PeterLG @jodmentum @sundogplanets

Even with a colider big as around whole Mars we never can have a stabile blackhole. The amount of energy needed only is not possible to get a STABLE blackhole in any way ever. Most confuse tiny unstable blackholes with a stable one. A stable one is physicaly not possible ever, at least not in a surviving way.

@PeterLG @ERROR_WTF @jodmentum @sundogplanets The moon datacenters would have all the same problems as datacenters in orbit, plus the moon dust, which is extremely abrasive and gets everywhere.

Plus the temperature difference between lunar day and lunar night, which is almost 300°C, and all the material contraction and expansion that causes. Datacenters in orbit would at least spend all their time in sunlight (for the solar power to work).

@ticho @PeterLG @jodmentum @sundogplanets Yes this was a bit shortminded by myself. It was just an idea, and @PeterLG conviced me yet, that it is no good idea. The wish was bigger than the reality.

@ERROR_WTF @PeterLG @jodmentum @sundogplanets I can't blame you, it does sound cool, especially to a sci-fi fan. :)

@ticho @PeterLG @jodmentum @sundogplanets It was also because of the concerns where todays politcs could bring the earth. I am not not aware of some physics ( https://github.com/error-wtf/Segmented-Spacetime-Mass-Projection-Unified-Results ) but in some form real backup scenarios of critical data and technology out of some reach of certain destruction that can acure, is still, even this shortminded idea was spilled too fast, a kind of need. It has not only to do with scify, but with real dangers of our time.

GitHub - error-wtf/Segmented-Spacetime-Mass-Projection-Unified-Results: This repository provides a full Python-based implementation and verification of the Segmented Spacetime Mass Projection Model, offering a high-precision, testable alternative to traditional gravitational models.

This repository provides a full Python-based implementation and verification of the **Segmented Spacetime Mass Projection Model**, offering a high-precision, testable alternative to traditional gra...

GitHub

@ticho @PeterLG @jodmentum @sundogplanets Even the first goal should be to conserve what we still have as living conditions on this planet, also we still only have this only planet, we should not have the dangers as a blind spot. If we want to survive as a species, we should still at least focus, on how to still have the time to prepare the brought of the earth lifeforms, in form of genomes, (we ourself never can), to another planet out of our solar system. At least we should save what we have.

@ticho @PeterLG @jodmentum @sundogplanets

If you have never failed, you can't know anything.

*LoL*

@ticho @PeterLG @jodmentum @sundogplanets

The Earth is not a testing ground – it is our only habitat.

Mars terraforming is tech porn for billionaires.

Backup systems or systems that can restore life on our planet, because our planet is the best compromise we have, or even prevent destruction scenarios in the first place, are desirable.

Just because we don't have a conclusive solutions yet, you should not stop thinking about the how maybe and what could provide this safety.

Now I stop...

MistaPlanet Feb 10

@jodmentum @sundogplanets

You started with a vague statement. Then proceeded to telling us about a totally unrelated Google/NASA, and then ended the post with the same vague statement you started with.

In everything you wrote, you didn't mention how it is a terrible thing to do or why you say it is unrealistic

Try this for size... 🤷‍♂️

https://theblower.au/@PeterLG/116044955361792005

@jodmentum @sundogplanets

Peter (@[email protected])

@[email protected] Sure, IF we could lift that much equipment to geosynchronous orbit, and IF we could construct large orbital infrastructure 22,000 km up, and IF we could get fast, effective, stable, high-speed, low-lag communications from ground to orbit, and IF we could maintain that infrastructure daily, and IF we could manage the heat dissipation problem, and ... it might make sense. Or once we are able to permanently live off-planet. There are too many ifs for our current level of space tech. Not to mention: in a nuclear war, ground infrastructure will be destroyed, and any space-based infrastructure would be sitting duck for satellite killers (a ballbearing with attitude). One day, if we can get our shite together as a planet, but not while big country A has designs on big country B for resource or ideological "needs". @[email protected] @[email protected]

The Blower

@mistaplanet @jodmentum @sundogplanets

Just read it again. Why won't datacenters in space nit wirk? There are issues with power, there are issues with heat management, with data transmission, with vulnerable equipment in a harsh environment. Still don't get it? Read it again.

MistaPlanet Feb 10

@astolk @jodmentum @sundogplanets

I just read the article now.
Quite brilliant

It's just another Musk Brain Fart® that employees have to waste time on instead of doing real science.

Sablebadger Feb 10

@jodmentum @sundogplanets "but.. but.. musk is a genius! He's smrt an stuf! Buy moar stock! All his ideas are gold!"

Just don't look to hard at whose succesful idea it was he stole and drove into the ground, or just how high he was on ketamine when he came with the other not-successful ideas he didn't steal.

He stole tesla, and his big contribution was the cybertruck. Space X I think he just got lucky with, and they probably keep him away from anything explosive.

Trump & Epstein: Biz partners.Feb 10

@Sablebadger @jodmentum @sundogplanets
I wouldn't say "got lucky" with SpaceX.

If NASA built rockets that exploded as often as his, they never would have landed a man on the moon.

#TheElongatedMuskrat put #DonnieDumbass in office b/c T**** is easily bought off (747's and fake awards) who will give you whatever you want for a price.

#Biden saw the danger of our entire space program dependent on an erratic nut like Musk and started courting Boeing. Thus, he had to go.

@jodmentum @sundogplanets

There's another big reason this isn't going to work and that's economics. How much does it cost to get stuff up there? Like dollars per rack? I guess you can buy several hundreds of racks, maybe thousands, for the price of getting one into orbit. All other technical problems aside, if it is more expensive on one side, are the benefits on the other side of the ledger like hundred times bigger?

Magnus Ahltorp Feb 10

@jodmentum @sundogplanets Elon Musk loves hoaxes that distract from actual solutions. He has no serious intention of actually making this.

See Hyperloop, where the only purpose was to distract from trains. We’ve known for at least 12 years that he is a total fraudster.

Chris Real Feb 10

@ahltorp @jodmentum @sundogplanets

Some of us have not forgotten . . . or forgotten about, the "water purifiers" that never showed up in Flint, MI, —MANY years later, the promise has never been kept.

But this con man cares about nothing besides his reputation, which is phoney . . .

Santiago, né ?

@_chris_real @ahltorp @jodmentum @sundogplanets Isn’t it fascinating how it always works ? “I am going to put brains in ravioli and they will serve you ! So just invest in my car company and one day it’ll happen “

Chris Real Feb 10

@santi @ahltorp @jodmentum @sundogplanets

It's not fascinating—it's how the distantly desperate push the buttons that keep them safe, while subtley snitching on the sources of corruption that are Too Pervy To face judgment.

If they get prosecuted, they will be more surprised than the people who corner them.

Not A Number Feb 10

@sundogplanets @jodmentum it's not supposed to make sense, it's just supposed to be a cover to bury a basket case business (X) inside a profitable one (spaceX).

@jodmentum I loved when a learned person with all the right qualifications steps up and point by point describes why the latest technofeudal idea is hot air. I will bookmark this article in my notes, so I don't need to read anything more about this, and can pull it up and share it every time I meet some unlucky person who got caught in the spell.

@jodmentum @sundogplanets

ICE detention centers in space? Who's with me? guys?

Talia Hussain Feb 10

@jodmentum @sundogplanets and yet, I still think sending all the data centres to space is a very good idea that we should implement immediately

@jodmentum @sundogplanets he worked there - nothing happened

Mad Engineering Feb 10

@sundogplanets @jodmentum I'd worry about energy, ionizing radiation, a propagation delay.

@jodmentum @sundogplanets one person's opinion is a terrible idea

Quark Maker Feb 10

@jodmentum @sundogplanets So, I propose we call this the "Great Space Data Race Boondoggle", and furthermore I propose that companies absolutely *should* do this anyhow, for *precisely* the reasons Taranis has put forward.

It would cause a wide ranging, catestrophic failure of any corporation involved, which would then put them out of the way so that other, better run companies could take over.

Trump & Epstein: Biz partners.Feb 10

@jodmentum @sundogplanets
I learned an awful lot from that article.

All the "obvious" advantages of putting an #AI #DataCenter in space simply aren't true.

The cold of space being used for cooling? No convection; no cooling. (Read article for details.)

Solar panels to generate the power? No more efficient than ground-based solar panels.

Free up land by moving to space? The electronics would have to be so large it would "dwarf the ISS."

A very good debunking.

Maria Langer | 📝💎🌵🛥️Feb 10

@jodmentum @JSteven @sundogplanets This is real? I thought it was a joke!

Hubert Figuière Feb 10

@jodmentum it's all about the IPO. Make some smoke and mirrors for the investors.

@jodmentum @sundogplanets i think there is 1 argument in favor of it and that is, that is really costly for law enforcement to confiscate harddrives in orbit and they certainly cannot do it without the observant owner noticing it in advance. i imagine it to be really hard to launch a rocket without being seen

Dr Susi Arnott Feb 10

@jodmentum @sundogplanets
Thanks for this:)
"the kind of electronics needed to make a datacenter work, particularly a datacenter deploying AI capacity in the form of GPUs and TPUs, is exactly the opposite of what works in space"

@jodmentum @sundogplanets what disturbs me most in all of that is that it doesn't seem to be obvious to anybody privileged enough to have any decision making power in any relevant environments, that at least any two of these four issues presented, are real show stoppers – at least as a very general idea without all details. ANYONE should have as much basic science knowledge to understand, that this won't be economically viable in the near future under any circumstances!

Wil Shipley Feb 19

@jodmentum @sundogplanets @gregtitus

It still amazes me how Elon and his ilk just ignore science and assume they know eveything.