Artificial Intelligence
Feb. 6th, 2026 11:08 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
ysabetwordsmith"Is AI more important than climate?"
When the BBC recently asked Google and Alphabet CEO Sundar Pichai whether the build-out of AI is more important than climate, the question briefly cut through the hype that usually surrounds the AI boom. Pichai acknowledged that AI is dramatically increasing energy in ways current systems “can’t fully cope.”
Another way in which humanity is too stupid to stop sawing off the branch we're all standing on.
AI is not more important than the climate, it is just the latest threat to the climate. AI is a massive energy hog that we cannot afford at a time when we need to be cutting emissions as fast as possible. The most effective way to do that is to use less energy. AI is the opposite of helpful in this regard.
In other words: when AI strains the system, the answer is to build a bigger system, not reduce the strain. As a self-defined technologist, he expressed confidence that “we will have abundant renewable energy in the future.”
In order to have even a chance at that working, we'd need solar panels as efficient as leaves. We're nowhere near that -- and already facing limitations due to scarce supplies of critical elements. Bluntly put, this is fantasy.
And Pichai made that worldview explicit when he warned governments, including the UK, that they “don’t want to constrain an economy based on energy,” because such limits “will have consequences.” In other words: economic growth must not be stalled, therefore energy supply must expand indefinitely to support it. This is precisely where technological optimism diverges from ecological reality.
Sadly a lot of people are divorced from reality nowadays.
Economist experts like Grace Blakeley warn the AI boom may be another bubble fuelled by hype and geopolitics, noting that when it bursts, it’s those who profited most will be shielded, while the public bears the cost.
Well, duh. It's mostly hype at this point; very little AI is actually profitable.
Right now, it is Earth’s life-support systems, upon which all markets depend, that are breaching their safe limits.
Grim as the situation is, I'm delighted to see someone else saying the kind of thing I've been pointing out for years: The economy cannot be in conflict with the environment, because the economy is a subset of the environment. No environment, no economy.
According to the 2025 Planetary Health Check report, humanity has already breached seven of the nine planetary boundaries, including climate, biodiversity, freshwater stability, land use and chemical pollution. These boundaries are not abstractions; they are the ecological foundations on which stable societies depend.
Well, that's horrifying. O_O
Google alone plans to invest $90 billion annually on AI-related infrastructure, as Sundar confirmed in the BBC interview. Across the industry, AI-related capital spending now exceeds $3 trillion.
The really frustrating thing is that most AI is anti-useful. Because it can't give sources, nothing it says is reliable and it should all just be ignored. So now search engines have a pile of crap at the top that people have to scroll past to find the information that is actually usable. Educational videos are increasingly inclined to have images in them that are just dead wrong. Employees forced to use AI at work report that it takes longer to correct the program's many fuckups than it would be just to do the work themselves like they used to do.
This is the largest private capital deployment in history, emerging in the exact decade scientists insist we must reduce global energy and material throughput to stay within ecological limits.
We are scaling the most energy-hungry computational paradigm humanity has ever built, precisely when the world can least support it.
Well said.
>>AI isn’t digital. It’s physical.<<
It is digital, but like all software, it is based on hardware, which is physical. Both aspects are important to keep in mind.
>>Data centres account for under 3% of global electricity use, yet this year it’s projected to rival that of Japan. By 2030, they could consume 1.2 trillion litres of water annually — enough for four million U.S. households — even as around a quarter of today’s data centres sit in regions facing water scarcity by 2050. According to Google’s own 2025 Sustainability Report, the company’s total water consumption rose 28% by nearly 1.8 billion gallons from 2023 to 2024.<<
We don't have that much water to spare either.
Solar panels and wind turbines last 20-30 years and depend on mining of finite materials,
IF you're lucky. Planned obsolescence is a thing that makes anything purported to have a long service life a very risky gamble.
Net-zero solves for carbon, not the conditions for life
An astute observation.
>> What scale of AI is compatible with a finite, destabilising planet? <<
Very little. We might manage to keep limited things from the past like spellchecking. We better manage search engines or the internet is fuckall useless. The most useful recent AI development I've heard of using it to spot target cells in medical samples, allegedly better than human experts; I would still recommend using both. But I don't know how much of an energy hog that is. We certainly can't afford all the generative AI that companies are slopping around.
>>Scaling energy- and material-intensive technologies while breaching multiple planetary boundaries is not innovation. It’s blind hope, disguised as progress.<<
I don't think it's hope. I think it's plain old greed.
While investors and executives push for unconstrained acceleration, my research with the LSE Post-Growth Transformation Lab found that the majority of UK professionals working in for-profit organisations believe businesses should operate within social foundations and ecological boundaries, in line with doughnut economics principles. In other words, the people building and running these systems already understand what leadership had yet to admit: limits are not anti-innovation — they are a condition for long-term business legitimacy and survival.
Good luck with that. It would certainly be an improvement.
What You Can Do
Let's look at those planetary boundaries...
* Ocean Acidification
5 Ways We Can Stop Ocean Acidification
* Freshwater Change
HOMEOWNER’S GUIDE TO: Protecting Water Quality and the Environment
* Land System Change
Helping Communities Protect Wildlife Habitat
Regenerative Agriculture
* Climate Change
50 Things You Can Do to Fight Climate Change (by categories)
* Modification of Biogeochemical Flows
Reducing Nutrient Loss: Science Shows What Works
* Introduction of Novel Entities
Local-Earth gengineering ethics are garbage. Frex, it is not ethical to create gengineered organisms that can get loose and contaminate wild or domestic stocks of natural genetics. Since people aren't using this technology responsibly, it is best to avoid as much as you can.
* Change in Biosphere Integrity (extinction rate)
Regardless of various people frantically trying to cover it up, the Anthropocene Extinction is very real. Just calling a spade a spade is helpful.
Facing Extinction: 9 Steps to Save Biodiversity
Understand Promiscuous Pollination in Landrace Gardening
The Buffalo Seed Company offers many landraces.
When the BBC recently asked Google and Alphabet CEO Sundar Pichai whether the build-out of AI is more important than climate, the question briefly cut through the hype that usually surrounds the AI boom. Pichai acknowledged that AI is dramatically increasing energy in ways current systems “can’t fully cope.”
Another way in which humanity is too stupid to stop sawing off the branch we're all standing on.
AI is not more important than the climate, it is just the latest threat to the climate. AI is a massive energy hog that we cannot afford at a time when we need to be cutting emissions as fast as possible. The most effective way to do that is to use less energy. AI is the opposite of helpful in this regard.
In other words: when AI strains the system, the answer is to build a bigger system, not reduce the strain. As a self-defined technologist, he expressed confidence that “we will have abundant renewable energy in the future.”
In order to have even a chance at that working, we'd need solar panels as efficient as leaves. We're nowhere near that -- and already facing limitations due to scarce supplies of critical elements. Bluntly put, this is fantasy.
And Pichai made that worldview explicit when he warned governments, including the UK, that they “don’t want to constrain an economy based on energy,” because such limits “will have consequences.” In other words: economic growth must not be stalled, therefore energy supply must expand indefinitely to support it. This is precisely where technological optimism diverges from ecological reality.
Sadly a lot of people are divorced from reality nowadays.
Economist experts like Grace Blakeley warn the AI boom may be another bubble fuelled by hype and geopolitics, noting that when it bursts, it’s those who profited most will be shielded, while the public bears the cost.
Well, duh. It's mostly hype at this point; very little AI is actually profitable.
Right now, it is Earth’s life-support systems, upon which all markets depend, that are breaching their safe limits.
Grim as the situation is, I'm delighted to see someone else saying the kind of thing I've been pointing out for years: The economy cannot be in conflict with the environment, because the economy is a subset of the environment. No environment, no economy.
According to the 2025 Planetary Health Check report, humanity has already breached seven of the nine planetary boundaries, including climate, biodiversity, freshwater stability, land use and chemical pollution. These boundaries are not abstractions; they are the ecological foundations on which stable societies depend.
Well, that's horrifying. O_O
Google alone plans to invest $90 billion annually on AI-related infrastructure, as Sundar confirmed in the BBC interview. Across the industry, AI-related capital spending now exceeds $3 trillion.
The really frustrating thing is that most AI is anti-useful. Because it can't give sources, nothing it says is reliable and it should all just be ignored. So now search engines have a pile of crap at the top that people have to scroll past to find the information that is actually usable. Educational videos are increasingly inclined to have images in them that are just dead wrong. Employees forced to use AI at work report that it takes longer to correct the program's many fuckups than it would be just to do the work themselves like they used to do.
This is the largest private capital deployment in history, emerging in the exact decade scientists insist we must reduce global energy and material throughput to stay within ecological limits.
We are scaling the most energy-hungry computational paradigm humanity has ever built, precisely when the world can least support it.
Well said.
>>AI isn’t digital. It’s physical.<<
It is digital, but like all software, it is based on hardware, which is physical. Both aspects are important to keep in mind.
>>Data centres account for under 3% of global electricity use, yet this year it’s projected to rival that of Japan. By 2030, they could consume 1.2 trillion litres of water annually — enough for four million U.S. households — even as around a quarter of today’s data centres sit in regions facing water scarcity by 2050. According to Google’s own 2025 Sustainability Report, the company’s total water consumption rose 28% by nearly 1.8 billion gallons from 2023 to 2024.<<
We don't have that much water to spare either.
Solar panels and wind turbines last 20-30 years and depend on mining of finite materials,
IF you're lucky. Planned obsolescence is a thing that makes anything purported to have a long service life a very risky gamble.
Net-zero solves for carbon, not the conditions for life
An astute observation.
>> What scale of AI is compatible with a finite, destabilising planet? <<
Very little. We might manage to keep limited things from the past like spellchecking. We better manage search engines or the internet is fuckall useless. The most useful recent AI development I've heard of using it to spot target cells in medical samples, allegedly better than human experts; I would still recommend using both. But I don't know how much of an energy hog that is. We certainly can't afford all the generative AI that companies are slopping around.
>>Scaling energy- and material-intensive technologies while breaching multiple planetary boundaries is not innovation. It’s blind hope, disguised as progress.<<
I don't think it's hope. I think it's plain old greed.
While investors and executives push for unconstrained acceleration, my research with the LSE Post-Growth Transformation Lab found that the majority of UK professionals working in for-profit organisations believe businesses should operate within social foundations and ecological boundaries, in line with doughnut economics principles. In other words, the people building and running these systems already understand what leadership had yet to admit: limits are not anti-innovation — they are a condition for long-term business legitimacy and survival.
Good luck with that. It would certainly be an improvement.
What You Can Do
Let's look at those planetary boundaries...
* Ocean Acidification
5 Ways We Can Stop Ocean Acidification
* Freshwater Change
HOMEOWNER’S GUIDE TO: Protecting Water Quality and the Environment
* Land System Change
Helping Communities Protect Wildlife Habitat
Regenerative Agriculture
* Climate Change
50 Things You Can Do to Fight Climate Change (by categories)
* Modification of Biogeochemical Flows
Reducing Nutrient Loss: Science Shows What Works
* Introduction of Novel Entities
Local-Earth gengineering ethics are garbage. Frex, it is not ethical to create gengineered organisms that can get loose and contaminate wild or domestic stocks of natural genetics. Since people aren't using this technology responsibly, it is best to avoid as much as you can.
* Change in Biosphere Integrity (extinction rate)
Regardless of various people frantically trying to cover it up, the Anthropocene Extinction is very real. Just calling a spade a spade is helpful.
Facing Extinction: 9 Steps to Save Biodiversity
Understand Promiscuous Pollination in Landrace Gardening
The Buffalo Seed Company offers many landraces.
