AIBU to point out that datacentres don't provide local employment?

[SadiraOfTyr]

The government appears to be making a big deal about datacentres (AI or not) being a huge growth area. The reality is that datacentres, once built, employ virtually no-one - a handful of security guards and that's it. Maintenance, repairs, cooling and power, racking and cabling of servers, storage, and networking kit is intermittent and done by specialists brought in for the job. Most of the time datacentres are a (very noisy) ghost town.

Datacentres are all managed and run remotely by SREs (site reliability engineers) who can be stationed hundreds of miles away, and often in a different country.

YABU: Datacentres will provide thousands of high quality jobs in neglected parts of the country long after construction.
YANBU: The government is having the wool pulled over its eyes by big tech companies who, even if all these datacentres are completed, will never employ more than a handful of locals in low pay occupations such as security and cleaning.

I'm willing to be proven wrong but data centres in space is something I can't get my head around. None of the big problems about terrestrial data centres - power, cooling, cost, bandwidth, the sheer volume of space they take up - are in any way reduced by putting them in orbit.

You make a point re employment once built. Taxes need to come from that data centre somehow not the people working as there wont be many.

I would be very interested to hear how Musk intendes to cool these satellites, and how he will address network latency and bandwidth. My bet is on them not happening, at least not in the form proposed (747-sized satellites equipped with vast folding ammonia-filled radiators).

And where do you get that 20% of the UK's GDP is dependent on space research??

So if he's right, what happens once we've cured cancer, and presumably all the other diseases too? And all the new diseases that pop up? We're just kept alive to get older and older, and feebler and feebler, requiring the least impaired to provide 24hr care to everyone else? Well, until the robots take over and rescue all of us from ever needing to leave the house at all, and from any physical contact with other humans, as our minds and bodies fail and we're kept hooked up to machines to keep our hearts pumping.

Or will it figure out how to keep ourselves mentally and physically healthy for all eternity, so we no longer sicken or fade in any way? The world might start looking a bit crowded pretty quickly... So babies won't really be an option, however desired they may be. Just an increasingly aged population with nothing to do, hoping their accidental death isn't too gory. Or becoming more and more terrified and depressed to the point they take matters into their own hands.

And this is the best-case scenario, Altman's utopia. I gotta say, I'm not sold, personally.

https://www.gov.uk/government/publications/factsheet-the-uk-space-sector/factsheet-the-uk-space-sector

18% rounded up is 20%. By definition if 18% of the UKs GDP relies on space, then that 18% relies on space research.

Or maybe we should just stop ?

@SerendipityJane

What that report actually says is that sectors that make up 18% of the UK's GDP are reliant on satellite services. Which is a slightly cheeky way of saying that lots of sectors use GPS and satellite-based telecommunications.

Power, cooling and space is the draw. The idea is they're powered by solar arrays, which as they can be facing the sun 24/7, without any sort of atmosphere or weather obscuring them. Which means they'd be vastly more powerful than any earth-bound solar panels. Cooling is helped enormously by the fact space is really, really cold. And in sci-fi dream world, the heat the centres produce will be used create power, specifically for extra-terrestrial bases such as a moon base. And space they take up is solved by space being really... well, spacious! But space they take up in space ships to get them there is definitely a bit of an issue. As is cost, and bandwidth. But honestly, I wouldn't bet against them. Just probably not on Musk's timeline.

But it wouldn't really help the UK as we have no spaceships, and no money. So we'd still be entirely reliant on America, and as presumably the point of building our own centres is to reduce our reliance on America, this isn't much of a solution. Though it could be argued what's the point of destroying our own very limited land for it when the tech will be obsolete by the time it's built. But at our building speed, that's pretty much a given regardless, space or no space.

They can always stop.

Otherwise the premise of my point stands. Research into space is going to underpin GDP. Which will be needed if there are fewer jobs.

Not sure why they need to stop? I agree, space research is very important. I personally believe that interplanetary expansion is, long-term, humankind's destiny.

But that has nothing to do with Musk's plans to launch 747-sized 'AI satellites' that will somehow magically keep cool (a problem only solveable on Earth with vast cooling infrastructure), while also overcoming the speed of light to provide as low latency as terrestrial datacentres.

@rumblegrumble it doesn't matter how cold space gets (and in Low Earth Orbit it is really not very cold). It is extremely hard to keep things cool in the vacuum of space because you have no convection - all your cooling has to be done by radiation. You therefore need huge radiators which are kept at right angles to the sun at all times.

Starmind's design is for 150kW peak power and 120kW continuous. That's the same as the International Space Station. A typical terrestrial datacentre is 50MW. So we're talking about launching 400+ ISSs, just to match the compute power of a single terrestrial AI datacentre.

There are surprisingly few orbits were you stay in the sun 24/7 as the earth gets in the way for nearly all of them. Typical satellites deal with that by relying on batteries while they're in the shade but the battery capacity limits their power load. Data centres need a lot of power.

Meanwhile, all the time your satellite is sitting in direct sunlight it's getting hit with over 1KW of radiant energy per square metre as there's no atmosphere to protect it. Only a small fraction of that energy can be usefully converted to electricity and the rest just makes your satellite hotter. Even worse, space is a vacuum and vacuums are really efficient insulators. That's why Thermos flasks keep your tea hot for so long and also why the International Space Station has absolutely enormous radiators to help it dump the heat and avoid cooking its occupants.

I’ve visited loads of data centres as part of my job and there are plenty of local people working at them. There are also a lot of ex military, security, engineer roles that may mean people move to an area, but as an example I have been to a huge data centre in Ireland and it felt about 80% local workers.

I’ve been to data centres in northern Sweden and in Denmark and the US and they were mostly locals (again electrical engineers, typical health and safety type roles) with maybe around 30-40% who had moved specifically for the roles.

@GasperyJacquesRoberts and all the radiation that you DO manage to convert into electrical power ends up as heat anyway.

Space construction is mainly engineering, really. Even the Apollo programme. Very few "can'ts" as long as you throw enough money at it.