| ▲ | usrnm 6 hours ago |
| From a purely engineering perspective, this and the destroyed AWS datacenters in the Middle East show an interesting trend. How will it affect the way future datacenters and services hosted in them are built? Will we see a rise in decentralized systems and much larger networks of smaller DCs? |
|
| ▲ | swiftcoder 5 hours ago | parent | next [-] |
| > Will we see a rise in decentralized systems and much larger networks of smaller DCs? Or will we see civilian datacenters with their own defensive drone swarms, and a couple of phalanx-style point defence systems for good measure? |
|
| ▲ | yablak 6 hours ago | parent | prev | next [-] |
| No, just more underground construction |
| |
|
| ▲ | pu_pe 5 hours ago | parent | prev | next [-] |
| This should have a material impact on SpaceX stock. Wouldn't it be much easier to harden datacenters against attack under ground than in space? |
| |
| ▲ | fch42 2 hours ago | parent | next [-] | | To merely destroy "space data centers", you need not have very advanced targeting technology nor space lasers or whatnot. Just launch a few things releasing little ball bearings or steel needles or something like that - and "poison" the entire orbit. If all you seek is to destroy (how en-vogue...) then that's more than perfect; not just destroy what's there but also make sure that by virtue of what remains (debris) nothing will get re-built. | | |
| ▲ | spiderfarmer 8 minutes ago | parent [-] | | A single EMP would be enough without causing excessive space debris. It's not like someone will simply fly in and swap some hardware. |
| |
| ▲ | adventured 3 hours ago | parent | prev [-] | | The SpaceX stock is floating solely (and temporarily) on fairy dust, it will not have any impact on the stock what-so-ever. As the market goes, so goes the multiple being granted to SPCX. That is, when the AI bubble contracts on the multiples being granted, SPCX will decline dramatically. In my observation HN typically widely can't differentiate properly between the real AI economy being developed and the multiples being temporarily granted in the stock market, it's common on here that people think they're the same thing ('when the AI bubble ends'). Back in reality what is going to happen is an implosion of the multiples when the growth slows (or suddenly stops due to an economic shock), and despite that the AI economy broadly will radically expand regardless over 10, 20, 30 years. See: the Internet build out (and market valuations), it will occur much the same conceptually. | | |
| ▲ | kolinko an hour ago | parent [-] | | SpaceX has a de facto monopoly on space now. And they are 10-20 years ahead of even China. No nation now has a response to Falcon 9 - sth SpaceX built over 10 years ago. |
|
|
|
| ▲ | childintime 5 hours ago | parent | prev [-] |
| It will mean data-centers move to space and then Musk will use his space prowess to bring weapons into space, in collaboration with the US military, and deny/strong-arm/corrupt governments around the world to protect his treasure, including the US government. He'll have a robot army to back it up. He is on his way to become the first ($100 trillion) supervillain. Until 2 years ago, he stood at the center of most of the good news, then he got a few calls from Putin and he became unhinged from our (obscured) point of view. Nowadays he's a dark pattern. Capitalism culminates in him, and he has to decide what to make of the world. He is our de-facto king, while on a personal level he has to decide what to do with his fortune. Put it into smart contracts for an AI to use and rule the world? Giving it to his kids would be underwhelming, not what he lived for. People joked that mankind existed to bring forth the machines and would then become superfluous. The irony of it. Well, it seems irony has mr Musk. He sees the petty online bickering and he sees himself as alone, as the only adult in the room, who has to decide the fate of humanity, and his power in many domains puts him to the task. He has his fiduciary responsibility that puts the interests of his share-holders first. He sees that as a healthy check on his power. At the end of all his calculations he has responsibilities akin to the Star Wars Emperor. Just like many white men, who come spend the remainder of their days in function and in defense of their property. Not in defense of life. So Musk needs his antagonist to arrive on the scene, so he can repurpose his empire for the good the antagonist brings, as he represents a complete overhaul of the way life is lived. So the jury is out on mr Musk, and may he always remember his status as an American is just a hustle gone wild. Similar to the dynamic of how a liar has to keep extending his web of lies. He will make it into a machine that will carry his burden going forward. And then what? When, if ever, will he remember he is an African, and a friend of his own antagonist? His final words may sound like: Luke, I am your Father! |
| |
| ▲ | TheOtherHobbes 4 hours ago | parent | next [-] | | Data centres will not be moving to space. The physics and economics don't work. It's not that it can't be done at all, but heat dissipation at scale in space is an insane engineering challenge, the tech to make giant radiators work doesn't exist yet, and even if it did it would extremely costly to develop and deploy. Musk is a textbook narcissist. His consistent pattern is to make grandiose promises - literal future faking - fulfil a small percentage of them, then make even bigger and more grandiose promises and hope investors have a short memory. (They do.) As a narcissist his primary relationship orientation is contempt for others. He's more interested in damaging projects he doesn't like - like the left-leaning community on the original Twitter, using promises of Hyperloop to kill high speed rail, ignoring the CSA from Grok, and of course the USAid fiasco - than seeing through the projects he claims to be developing to completion. He fits in perfectly with Thiel, Altman, and the rest. The tech is sometimes interesting, but the alignment is fundamentally hostile to the rest of humanity. | | |
| ▲ | ben_w 4 hours ago | parent | next [-] | | > but heat dissipation at scale in space is an insane engineering challenge, the tech to make giant radiators work doesn't exist yet, and even if it did it would extremely costly to develop and deploy Different problem. Cooling can be done in space, it's a normal problem; not easy, nothing in space is easy, but on the scale of space it is normal. However, at the temperature you would run a chip at, a radiator you make for space will just work better, reject more heat, if you leave it on the ground. We get convective cooling down here as well as radiative. At operational temperature, convection more than makes up for the atmosphere being a heat bath at ~20 C or whatever. | | |
| ▲ | spiderfarmer 5 minutes ago | parent [-] | | Everything is possible. Doesn't mean you should do it. You can also put datacenters on the sea floor in a nuclear submarine. That might be even easier. But does it make sense to do it? No. |
| |
| ▲ | Marha01 4 hours ago | parent | prev [-] | | > t's not that it can't be done at all, but heat dissipation at scale in space is an insane engineering challenge, the tech to make giant radiators work doesn't exist yet, and even if it did it would extremely costly to develop and deploy. You are overstating the engineering difficulty. Starlink constellation already dissipates around 100MW. | | |
| ▲ | fch42 an hour ago | parent [-] | | If you solve your cooling problem in space by distributing 10000 servers across 100 orbits then you replace one engineering challenge with the next; good luck with the satellite-satellite optical crossbars, and even more with the time travel beam shapers to get the communication latency between the components of your "orbit-distributed data center" to-target. (the sarcasm is intended; there are some usecases for "space servers", around stealth for sure; to "scale it like a terrestrial DC" though ... dunno; not matter how I turn it, it just seems crazy expensive) | | |
| ▲ | Marha01 32 minutes ago | parent [-] | | Communication latency and bandwidth would present big issues for AI satellites used for training. But if they are inference-only, then the latency and bandwidth demands are much, much lower. Starlink already demonstrates the required parameters. | | |
| ▲ | fch42 a few seconds ago | parent [-] | | Indeed ... yet the servers for that already exist and are largely ground-based - "edge" compute of the large CDNs through which much of the internet's traffic flows anyway. Will starlink build its own edge ? Maybe. We'll see. Whether it makes sense to have this infrastructure "colocated in space" if you're not also the owner of the "connection infrastructure in space" though? It'll come down to cost.
If I were anyone else but SpaceX ... that's where my eyes would be. |
|
|
|
| |
| ▲ | ben_w 4 hours ago | parent | prev [-] | | In 2016, Musk was forecasting no-asterisk-needed cross country self driving and Mars launches within two years. He may have been a wrong'un before that, but he was definitely already doing a confidence trick by that point. |
|
