Honestly the conversation of AI researchers on climate science seems so superficial and removed from reality of climate change. It's like they are trying to solve a drought by capturing condensation on a glass of cold beer.
The major emissions sectors are agriculture, transportation, cement and metal, energy usage in primary sector for thermal processes.
For someone who seems to have dedicated their lives to climate and AI solving energy intensity of deep learning isn't even relevant let alone a factor in emissions, also better climate models aren't the solution for reducing emissions.
We need not only new tech but a new socio-technical system for reduction of emissions. Rice cultivation, cement, waste, heating of houses in northern latitudes all generate sizeable quantum of emissions what's AI going to do there to solve these?
Optimization can squeeze some gains but we need revolution in fundamental science and engineering capabilities to quickly discover, synthesize and scale the manufacturing of new materials whilst rewiring the every elememt of the the supply chain to support such transition. This is a problem for AI- yes, but is this where bulk of climate AI research is focused on? Sadly not!
What is more motivating and upbeat though is the fact scientists from traditional engineering domains are increasingly discovering AI. SciML work by Chris rackauckas PiNN by karniadikis, neural operators by anima Nvidia are some of the best forays into this space. I have not mentioned the frentic research happening in AI and microbiology which can be a game changer in invention of biotechnology solutions for climate change.
Yet sadly none of this is ever mentioned or covered in climate AI related conversations.
If so, that would mean a policy choice of governments of these regions to build out the infrastructure to pipe the heat from the data centers to peoples homes, and to power those data centers with green energy, etc.
It is still not AI helping with climate change, but policy makers making sane policy choices.
I’m slightly confused. Climate scientists have been using machine learning to make their climate models since before it was called machine learning. AI has been at the forefront of fighting climate change for a few decades now. It has in fact helped shape policy, for governments to set targets (which get promptly ignored by politicians), etc.
Using AI to slightly optimize any individual’s carbon footprint in computation or by slightly optimizing energy distribution (as if it isn’t already pretty optimized) seems very minor compared to grand policy changes to build out green infrastructure and international agreements—which climate scientists are actually calling for based on the climate models which uses AI extensively.
The hype has also been great at fueling economic activity which contributes even more across a wide range of sectors.
We know how to reduce carbon emissions immediately, we did a great job of it 3 years ago, we just don't like the answer: reduce economic activity.
Consume less, produce less, work less, spend less. These things will do more to reduce CO2 emissions than if everyone rushed out and bought electric cars (in the US ~oil~ gasoline consumption has been positively correlated with electric car ownership).
But we don't want that, and more importantly, the stock market doesn't want that. I think most people would be fine with reduce work and reduced consumption (to a point), but the people invested deeply in the status quo tend to have asymmetric influence.
Even if we stopped all carbon emissions today the earth would continue to warm for centuries. Mitigating these effects is going to take carbon capture(aka unburning oil) and geoengineering on an immense scale which won't be possible without economic growth. There's no going back.
More importantly, the sooner we can stop carbon emissions the less severe warming and ecosystem harm we'll experience. Carbon capture may help a bit if we are able to massively scale it up, but there really is not substitute for ending emissions.
we have a choice to reduce economic activity in the controlled or uncontrolled way. sadly it seems we have to have tons of unrepairable electronics replaced every year and downfall of the civilization or it's "communism" aka ultimate evil
"Consume less, produce less, work less, spend less"
I want that. I'd love to trade money for time but the relationship seems non-linear. Working 50% as many hours would mean less than 50% as much pay for the most part.
IMO key is the cost of shelter. In all the desirable locations with good jobs the cost of accommodation rises so much that even well earning middle class in uncertain of their future. There is a number of key services and goods everyone needs: somewhere warm to sleep, some food, and healthcare. If people weren't constantly worried that they will lack this the whole nonsensical piramid of useless work would fall apart and everyone would be much happier
Very much depends on your country and your particular tax situation, of course. The US is pretty good for FIRE (aside from dealing with health insurance). Ireland (where I live) is terrible, by comparison.
this response ignores large and substantial inputs to a complex system of systems, instead choosing to focus on a one-line and highly contentious assertion that directly and falsely merges "growth producing carbon waste" with "economic output." Lazy oversimplification like the parent comment, is directly counterproductive to addressing serious and immediate design and political challenges.
Please consider other economic inputs that do not directly increase carbon emissions, and include design consideration where "new machines and methods" can accomplish similar or better economic activity.
That's not correlated with EV usage either. EV's comprise well under 1% of the cars in America, and went up during the pandemic. Your numbers are changing a lot more than 1% and went down during the pandemic.
Reducing carbon emissions isn't a solution. We need to go to all the way to zero and then to start carbon capture.
That ONLY happens with batteries fusion and AI. We are making progress and as long as the pessimists don't get their way.
Ironically people like Greta have had the worst impact on climate. Not just that she flies around the world instead of doing zooms. But that she is all doom and gloom instead of trying to help us progress technology
> But that she [Greta Thunberg] is all doom and gloom instead of trying to help us progress technology
I’ve been hearing this a lot lately, that climate doomerism is somehow damaging to the cause and preventing any meaningful climate action.
I think this is a misunderstanding of what climate doomerism is. I hear people talking as if being realistic about damning nature of the damage we have done, and who is to blame for it all. And instead people believe the message is that we should just give up. If you believe the latter, I dare you to go find any source where Greta Thunberg (or any other supposed climate doomer) has said anything of that nature (hint they haven’t).
But suppose that you are right, and the message of climate doomers is actually that of giving up, now if you compare that to the decades of misdirection and outright lying that the fossil fuel industry has done to actively hinder climate action, and on top of that to actually double down on burning fossil fuels. How much damage, relative to the fossil fuel industry, have people like Greta Thunberg actually caused. Relatively speaking that damage is a drop in the ocean.
I'm not sure what could be done much better than what Greta does given her environment. I believe so far her impact on climate awareness is large net positive, and that alone - the putting the problem front and center before many people - should already help at least somewhat with mitigating it. Not to mention other, largely positive effects of her work.
> We are making progress and as long as the pessimists don't get their way.
I've been hearing a similar statement for decades now, but the global emissions only continue to rise.
In fact, I've been hearing that logic since the late 90s and in that time we have doubled the cumulative emissions up to that point.
We have seen that emissions drop with decrease in economic activity (multiple times). Yes, batteries, renewables, and nuclear would all part a big part in a truly sustainable future... but we have to survive until that future first.
> Reducing carbon emissions isn't a solution
We're well past the "solution" stage of this problem. Where in the "harm reduction" phase. Reduced emissions is absolutely superior in this regard to increased emissions... and yet you're arguing that we should be increasing them because a magic solution is just around the corner, like it always has been.
Perhaps I am just optimistic but AI doesn't seem necessary, nor does fusion. Fission is more than capable of providing enough power and if we are going to do this via sequestration, we may as well start now and forget net-zero.
I definitely think this is an organisation problem, not a technology problem. Societies are not organising to fix the issue, instead it seems they are hoping the global free market will eventually make it economically beneficial in the short term, to save the climate.
There is plenty of data that suggests societies respond to financial incentives.
As the cost of battery/wind/solar plunge, their use goes up. And fusion has the potential to be much cheaper than those.
A very very small percentage of CO2 emissions is due to people wanting to increase CO2 in the atmosphere. Almost all of it is because that is the cheapest or most efficient way to travel, heat your home, make products, etc...
> Ironically people like Greta have had the worst impact on climate. Not just that she flies around the world instead of doing zooms. But that she is all doom and gloom instead of trying to help us progress technology
I'm pretty sure you're wrong here, and Greta doesn't fly, but if you insist, can you share your sources?
Greta and her PR team do not publish details of her carbon footprint, so I don't have exact details her travels but Wikipedia is a fairly reliable source to give some insight.
The word "emergency" occurs once, in phrase "Thunberg does not fly except for emergency cases.(4)" . The link (4) leads to https://time.com/5663534/greta-thunberg-arrives-sail-atlanti... "Climate Activist Greta Thunberg, 16, Arrives in New York After Sailing Across the Atlantic" . I didn't read the whole article, are you saying that article says Thunberg flies?
Sorry, I'm not sure where the wikipedia Editor got that information from. I can change my original comment to "travels with a significant carbon footprint" if you would make you happy? Does the carbon footprint section of wikipedia seem up to your standards?
From what I know about Greta, she is rather careful with such matters. Your comment in general is in effect destroying part of those careful efforts, so I'd be happy if it wouldn't.
"Significant carbon footprint" asks for, significant compared to what? Maybe Greta leaves bigger footprint than the lowest, say, 10% of the Earth population, who don't really have access to modern technologies with that significant footprint. But she surely tries quite a lot given the circumstances where she is, and encourages us to follow. So I'm skeptical about accusations towards her without enough evidence.
"Careful" isn't a number though. The wikipedia referenced flights to support her boat trip alone put her carbon footprint above hundreds of millions of people.
She is most likely in the 90% percentile not the 10% percentile. Which is my point. We shouldn't look to people with huge carbon footprints as role models. Al Gore for example has multiple homes. He is easily in the top 5% of carbon footprints in the world.
If the most alarmist and upset climate activists cannot get their carbon footprint down below the average human, then it is pointless to try. The only fix for this problem is technological acceleration toward clean energy, even if getting there causes short term emissions to increase.
I disagree, but I don't think this is a good platform to list arguments and supporting explanations. I've tried to write a longer post, but explanations are too long to be sufficient.
This article's first half was a general discourse on ML & the latter half seemed to revolve around how chip thermals & data center consumption can be managed by ML.
ML's contribution to reducing human-generated impact on climate isn't the prime focus of this talk. And that carbon footprint management is (as you might guess) miniscule in grand scale.
Seems like AI may contribute to emissions with all of the processing that has to be done gathering, formatting, training, and executing these models. Of course, in the long run if AI replaces most of our jobs, the population may drop, reducing emissions.
I guess, "help" here means "work towards achieving something desirable". Unless, of course, you think any and all kinds of capitalism regulations can't be defined so that they could be desirable...
https://ourworldindata.org/emissions-by-sector
The major emissions sectors are agriculture, transportation, cement and metal, energy usage in primary sector for thermal processes.
For someone who seems to have dedicated their lives to climate and AI solving energy intensity of deep learning isn't even relevant let alone a factor in emissions, also better climate models aren't the solution for reducing emissions.
We need not only new tech but a new socio-technical system for reduction of emissions. Rice cultivation, cement, waste, heating of houses in northern latitudes all generate sizeable quantum of emissions what's AI going to do there to solve these?
Optimization can squeeze some gains but we need revolution in fundamental science and engineering capabilities to quickly discover, synthesize and scale the manufacturing of new materials whilst rewiring the every elememt of the the supply chain to support such transition. This is a problem for AI- yes, but is this where bulk of climate AI research is focused on? Sadly not!
What is more motivating and upbeat though is the fact scientists from traditional engineering domains are increasingly discovering AI. SciML work by Chris rackauckas PiNN by karniadikis, neural operators by anima Nvidia are some of the best forays into this space. I have not mentioned the frentic research happening in AI and microbiology which can be a game changer in invention of biotechnology solutions for climate change.
Yet sadly none of this is ever mentioned or covered in climate AI related conversations.