I think we have two groups that incongruously oppose, while crossing political lines. One group see the technology as an existential threat to their careers, while the other sees little value in the technology.
I’ve seen executives extol its virtue in summarizing, but you can feel the pressure that is experienced by those being summarized. After all what if they just gave the raw data to the thing and it wrote the report, which it then summarized.
The skeptics ask it questions it cannot answer, either because the question was unclear (to a computer), or simply too complex for what is now a primitive application.
Neither camp sees value.
Agree about the article. Being ahead of the curve is kinda Elon’s superpower. And, yes, I think data centers in space will bypass at least the purely NIMBY aspects of the resistance. What resistance remains for non-NIMBY reasons would have to be confronted in any case.
It is a part of the cultural war against the west to bring it down and replace it with a communist utopia.
The next objection will be the carbon emissions of Elon’s rockets to launch space data centers.
The neo-Bolsheviks will certainly try to saddle and ride this particular pony, but they try that with everything. I suspect most of the people involved in data center opposition are not particularly ripe for conversion to Marxism and I don’t think the neo-Bolshies will be numerous enough in the ranks to succeed in bending this trend to their desired ends – though I would be amazed if they didn’t try.
My advice to anyone looking to build data centers in the US is to look at the place I was born and raised, Michigan’s Upper Peninsula. It’s big, thinly populated, has plenty of water available, has a well-established natural gas distribution network to feed in situ new-construction power plants and the pervasive forest is mostly second-growth or newer so “pristine-ness” is not much of an issue. And there are large stretches of it that are not particularly scenic. The politics of the rest of the state might be an issue, but the U.P. has tilted Republican forever.
I’ve already called the problem, it’s going to be NOx, not CO2. CO2 and Elon’s most robust projection is a rounding error compared to general aviation.
I am going to work up a more NOx-intensive piece in the future, after I get a few more articles out of my craw.
SpaceX (and Blue Origin and others) needs to be getting out in front of this.
I say this only half in jest, Elon needs to call me. Or more specifically, someone in the right division of SpaceX needs to have a conversation with me so I can make them aware of my findings before the luddite progressives (but I repeat myself) discover the issue and use it to kugel them.
In terms of community development, data centers are a complete bust. They consume massive amounts of power and enough water for a small to medium farm for very few, just so so, long term jobs. Much better for the community to spend their money and resources on some sort of manufacturing facility if they can find one. The actual location of the centers matters very little, put them where it’s cheapest. The people making the real money from the data centers won’t be living and spending their money anywhere near them.
As a confirmed AI skeptic, I don’t want any of my money or my tax money subsidizing something I suspect will turn into a massive financial disaster. I’ll start to take it seriously when I see somebody making an actual profit from it instead of giving billions of dollars of compute away to build “presence”.
I’m against tax money subsidizing almost anything. Very few things actually have to be government supported, and Datacenters, stadiums, and car companies shouldn’t be on the list.
Obstruction is deadly to progress and especially innovation. There are dozens of things that I can think of that could various situations if it were legal.
Housing for homeless would properly be tiny and Spartan. cheap daily rentals with Bubba and Leroy to enforce collections and vandalism. Zero chance of permitting. Or of being able to hire them at wages appropriate to their character.
Schools promoting by stuff learned instead of time served.
Two is enough as none of us have days for useless lists.
The problem with giving houses to the homeless is that they are incapable of living in them, especially if there are any rules.
Tiny houses are a common idea and somehow government manages to screw them up by making residents use a common kitchen and bathroom while spending millions per unit.
Not government tiny houses. Private apartment types that are profit only. Bubba and Leroy will beat the crap out of you for overstaying, vandalizing, or any other violation of decent existence. One strike and you are out. Not politically or legally feasible. Just something that would work for any legitimate homeless down on their luck. The ones that refuse to act responsibly at all can freeze. The idea is not politically correct but could function in reality if allowed. Say $10.00 a night in advance and no subsidies.
“Water consumption” for a data center is mostly for cooling. So it takes in cold water and releases warm water.
The water is consumed by evaporation in cooling towers absent close proximity to a large body of water, lake river or ocean, in which case thermal pollution is an issue. Think of the fish.
The opposition to orbital data centers is driven by the academia-led astrophotography people who constantly post long-exposure photographs “proving” satellite constellations will obliterate the night sky. A had a friend of mine who bought into that guff stand outside in a remote location at dusk and asked him to see if there were any streaks in the sky, then to count how many satellite he saw. In one hour, zero streaks (once he understood about meteors) and three moving-dot satellites. “But I can’t imagine these reputable astronomers are lying. You must be wrong.”
Note also that those satellites are only visible in those photos for a limited amount of time before sunrise and after sunset, when they are illuminated but the ground below isn’t. In my experience, one doesn’t see many satellites in the middle of the night.
(On several occasions, near dawn, I’ve seen the train of satellites from a recent SpaceX launch, and it’s actually pretty spectacular. Years ago it would’ve given rise to UFO reports.)
I think it’s a combination of thee factors: over-the-top marketing, failing companies grabbing the AI life preserver (routinely misinterpreted as “AI taking our jobs”), and an anti-rich narrative. For example, there have been repeated dire warnings, calls for temporary halts and regulations, etc. from the AI companies themselves. It’s a bizarre dynamic, they hype how dangerous their technology is in order to generate sales. Why are people to expect that AI is all wine and roses when AI businesses push doom and gloom so hard? That’s a self-inflicted injury.
Then there’s the companies on the way out. They’re trying any trick they can in order to stay above water. AI is seen as yet another tool of last resort.
I repeatedly hear of stories where low valued workers are replaced in schemes that just don’t make sense. For example, in a conversation, one poster described their high tech employer getting rid of interns by replacing them with AI. The interns used to write small programs for use by the professional senior programmers on the team. From a short term point of view, it’s a money loser. It’s probably cheaper for the senior coder to write the program directly rather than have an intern write up a mess and then have it debugged (and heavily fixed) by the senior coder. The long term point to this exercise was to develop the next generation of programmers in a low risk trial employment.
So they got rid of the interns. The interpretation was that AI was taking our jobs. But the company wouldn’t have done that if it were still thinking long term. They still have a need for a low risk hiring pipeline. And they’re getting rid of that pipeline while simultaneously shoving the work that the interns were doing on the highly paid senior coders. That’s not what a healthy company does.
Then there’s stories about job losses to AI. The story describes several artist types who allegedly lost work to AI. Reading between the lines, it’s a combination of failing companies doing stupid stuff (like firing a graphic designer and thrusting his workload on everyone else) and shady companies doing shady stuff (like generating AI art aping the style of an artist’s online work). Neither are serious businesses.
Finally, there’s the US political angle: billionaire envy works against Trump so it’s going to be pushed hard. And there are plenty of unpopular billionaires in the AI world (Musk, Bezos, Zuckerberg, etc.). It’s an easy move to make and it just increases the unpopularity of AI and data centers even further.
It was easy to see why datacenters would move to space. I do wonder what this means for all of those small companies that were going to do data processing in space. An excess in compute in a constellation could put them out of business.
It’s all just marketing. Rename it a cancer research center and call it a day.
I think we have two groups that incongruously oppose, while crossing political lines. One group see the technology as an existential threat to their careers, while the other sees little value in the technology.
I’ve seen executives extol its virtue in summarizing, but you can feel the pressure that is experienced by those being summarized. After all what if they just gave the raw data to the thing and it wrote the report, which it then summarized.
The skeptics ask it questions it cannot answer, either because the question was unclear (to a computer), or simply too complex for what is now a primitive application.
Neither camp sees value.
Very good article. I’d suggest even recommended reading for anyone considering the issue.
It seems Elon is ahead of the curve on this one – launching datacentres into space will bypass a lot of the rising resistance.
Agree about the article. Being ahead of the curve is kinda Elon’s superpower. And, yes, I think data centers in space will bypass at least the purely NIMBY aspects of the resistance. What resistance remains for non-NIMBY reasons would have to be confronted in any case.
It is a part of the cultural war against the west to bring it down and replace it with a communist utopia.
The next objection will be the carbon emissions of Elon’s rockets to launch space data centers.
The neo-Bolsheviks will certainly try to saddle and ride this particular pony, but they try that with everything. I suspect most of the people involved in data center opposition are not particularly ripe for conversion to Marxism and I don’t think the neo-Bolshies will be numerous enough in the ranks to succeed in bending this trend to their desired ends – though I would be amazed if they didn’t try.
My advice to anyone looking to build data centers in the US is to look at the place I was born and raised, Michigan’s Upper Peninsula. It’s big, thinly populated, has plenty of water available, has a well-established natural gas distribution network to feed in situ new-construction power plants and the pervasive forest is mostly second-growth or newer so “pristine-ness” is not much of an issue. And there are large stretches of it that are not particularly scenic. The politics of the rest of the state might be an issue, but the U.P. has tilted Republican forever.
I’ve already called the problem, it’s going to be NOx, not CO2. CO2 and Elon’s most robust projection is a rounding error compared to general aviation.
https://www.thespacereview.com/article/5226/1
I am going to work up a more NOx-intensive piece in the future, after I get a few more articles out of my craw.
SpaceX (and Blue Origin and others) needs to be getting out in front of this.
I say this only half in jest, Elon needs to call me. Or more specifically, someone in the right division of SpaceX needs to have a conversation with me so I can make them aware of my findings before the luddite progressives (but I repeat myself) discover the issue and use it to kugel them.
Unless Elon just wants to call me…
Let me endorse your call for a call.
And I think you meant cudgel there in the penultimate sentence. I wish someone would kugel me. Yummy!
Indeed, they have have already been doing that, so expect it to get more attention
In terms of community development, data centers are a complete bust. They consume massive amounts of power and enough water for a small to medium farm for very few, just so so, long term jobs. Much better for the community to spend their money and resources on some sort of manufacturing facility if they can find one. The actual location of the centers matters very little, put them where it’s cheapest. The people making the real money from the data centers won’t be living and spending their money anywhere near them.
As a confirmed AI skeptic, I don’t want any of my money or my tax money subsidizing something I suspect will turn into a massive financial disaster. I’ll start to take it seriously when I see somebody making an actual profit from it instead of giving billions of dollars of compute away to build “presence”.
I’m against tax money subsidizing almost anything. Very few things actually have to be government supported, and Datacenters, stadiums, and car companies shouldn’t be on the list.
No argument here. But governments should also not be unreasonably obstructive either.
Obstruction is deadly to progress and especially innovation. There are dozens of things that I can think of that could various situations if it were legal.
Housing for homeless would properly be tiny and Spartan. cheap daily rentals with Bubba and Leroy to enforce collections and vandalism. Zero chance of permitting. Or of being able to hire them at wages appropriate to their character.
Schools promoting by stuff learned instead of time served.
Two is enough as none of us have days for useless lists.
The problem with giving houses to the homeless is that they are incapable of living in them, especially if there are any rules.
Tiny houses are a common idea and somehow government manages to screw them up by making residents use a common kitchen and bathroom while spending millions per unit.
Not government tiny houses. Private apartment types that are profit only. Bubba and Leroy will beat the crap out of you for overstaying, vandalizing, or any other violation of decent existence. One strike and you are out. Not politically or legally feasible. Just something that would work for any legitimate homeless down on their luck. The ones that refuse to act responsibly at all can freeze. The idea is not politically correct but could function in reality if allowed. Say $10.00 a night in advance and no subsidies.
“Water consumption” for a data center is mostly for cooling. So it takes in cold water and releases warm water.
This isn’t sewage being released. They’re lying.
Of course they’re lying. But only on days whose names end in ‘day.’
The water is consumed by evaporation in cooling towers absent close proximity to a large body of water, lake river or ocean, in which case thermal pollution is an issue. Think of the fish.
An alternative, assuming it actually works out:
https://interestingengineering.com/military/saronic-to-build-4000-acre-texas-shipyard
The opposition to orbital data centers is driven by the academia-led astrophotography people who constantly post long-exposure photographs “proving” satellite constellations will obliterate the night sky. A had a friend of mine who bought into that guff stand outside in a remote location at dusk and asked him to see if there were any streaks in the sky, then to count how many satellite he saw. In one hour, zero streaks (once he understood about meteors) and three moving-dot satellites. “But I can’t imagine these reputable astronomers are lying. You must be wrong.”
Note also that those satellites are only visible in those photos for a limited amount of time before sunrise and after sunset, when they are illuminated but the ground below isn’t. In my experience, one doesn’t see many satellites in the middle of the night.
(On several occasions, near dawn, I’ve seen the train of satellites from a recent SpaceX launch, and it’s actually pretty spectacular. Years ago it would’ve given rise to UFO reports.)
Conventional microelectronics won’t survive the radiation in orbit. It will take something completely new from top to bottom in the way of hardware.
I get the feeling that the “dark fiber” boom and bust at the end of the last century is being repeated with “data centers”.
I think it’s a combination of thee factors: over-the-top marketing, failing companies grabbing the AI life preserver (routinely misinterpreted as “AI taking our jobs”), and an anti-rich narrative. For example, there have been repeated dire warnings, calls for temporary halts and regulations, etc. from the AI companies themselves. It’s a bizarre dynamic, they hype how dangerous their technology is in order to generate sales. Why are people to expect that AI is all wine and roses when AI businesses push doom and gloom so hard? That’s a self-inflicted injury.
Then there’s the companies on the way out. They’re trying any trick they can in order to stay above water. AI is seen as yet another tool of last resort.
I repeatedly hear of stories where low valued workers are replaced in schemes that just don’t make sense. For example, in a conversation, one poster described their high tech employer getting rid of interns by replacing them with AI. The interns used to write small programs for use by the professional senior programmers on the team. From a short term point of view, it’s a money loser. It’s probably cheaper for the senior coder to write the program directly rather than have an intern write up a mess and then have it debugged (and heavily fixed) by the senior coder. The long term point to this exercise was to develop the next generation of programmers in a low risk trial employment.
So they got rid of the interns. The interpretation was that AI was taking our jobs. But the company wouldn’t have done that if it were still thinking long term. They still have a need for a low risk hiring pipeline. And they’re getting rid of that pipeline while simultaneously shoving the work that the interns were doing on the highly paid senior coders. That’s not what a healthy company does.
Then there’s stories about job losses to AI. The story describes several artist types who allegedly lost work to AI. Reading between the lines, it’s a combination of failing companies doing stupid stuff (like firing a graphic designer and thrusting his workload on everyone else) and shady companies doing shady stuff (like generating AI art aping the style of an artist’s online work). Neither are serious businesses.
Finally, there’s the US political angle: billionaire envy works against Trump so it’s going to be pushed hard. And there are plenty of unpopular billionaires in the AI world (Musk, Bezos, Zuckerberg, etc.). It’s an easy move to make and it just increases the unpopularity of AI and data centers even further.
My first reaction to the NY state ban was: Governor, thank you for accelerating demand for space-based AI centers.
It was easy to see why datacenters would move to space. I do wonder what this means for all of those small companies that were going to do data processing in space. An excess in compute in a constellation could put them out of business.