16 April 2024. AI, again | Manure
The AI arguments are arguments about power. // Turning methane into an asset [#561]
Welcome to Just Two Things, which I try to publish three days a week. Some links may also appear on my blog from time to time. Links to the main articles are in cross-heads as well as the story. A reminder that if you don’t see Just Two Things in your inbox, it might have been routed to your spam filter. Comments are open.
1: The arguments about AI are about power
At his Substack the novelist and futurists Karl Schroeder has been reflecting on the amount of noise discussion of AI (and the spectre of AGI, Artificial General Intelligence, that sits behind it) is generating among futurists, as evidenced by the Association of Professional Futurists’ listserv and other places:
I regularly get called on to defend one or another position on the issue. I’ve recently done Delphi polls for policy recommendations, given talks, and been added to lots of listservs by people whose ideas and opinions I value highly. I really, really don’t want to be in this position.
I’ve watched those APF discussions myself, since I’m also on that listserv, and I know exactly what he means. But this also tells us something about the discourse about AI at the moment, so stay with it.
Schroeder is a science-fiction writer who has written about AI in multiple stories and books. He’s quick to make it clear that he’s familiar enough with the technology without being a technical expert, but what he’s bringing to the subject is an expertise in “mythopoetic structures”.
The current mythopoetic structure for the AI discussion goes like this:
We fear AI will become the servant that the Sorcerer’s Apprentice creates (as embodied in the famous paperclip maximizer). It’ll follow an instruction relentlessly, destroying everything and anyone in its way to achieving its goal.
The AGI discussion is like this, but more extreme:
AGI, on the other hand, is ... pitiless, tireless, and inimical. The mythopoetic archetype of such a being is the devil—or, which could be worse, the incomprehensible and arbitrary Greek or Norse gods.
(The Sorcerers’ Apprentice. Image by mofurgi/ Deviant Art. CC.)
From this he goes into a discussion of current competing views of what it means to be a thinking being. The first of these is the idea of ‘utility functions’, from American analytic philosophy and classical economics.
The second is shorthanded as ‘4E’, and says that cognition is “Embodied, Enacted, Extended, and Embodied” (his emphasis). He calls this one ‘Enactivism’. This has its roots in ideas such as autopoesis.
Under the first of these,
thinking beings are rational actors attempting to maximize some quantity that represents an objectively determinate value of reward versus effort... The mythopoetic structure of the utility-function AGI is that is a version of the disembodied “view from nowhere” rational mind of Newtonian science. As such, it hovers above the world, reaching in and moving pawns and knights across the board to achieve its ends.
Under the second, in contrast,
In order to continue to exist, an embodied and embedded entity may have to change not just is environment, but itself. The enactivist view is that organism and environment make up a complementary pair, each in some senses giving rise to the other. ... The enactivist framing of AGI would be of a paired system, AGI+environment, whose two halves coevolve.
Put like this, the arguments about both AI and AGI are a version of arguments about the history of science: Newton versus quantum. But the other thing they are about is about power.
What’s lying underneath all our anxieties about AGI is an anxiety that has nothing to do with Artificial Intelligence. Instead, it’s a manifestation of our growing awareness that our world is being stolen from under us. Last year’s estimate put the amount of wealth currently being transferred from the people who made it to an idle billionaire class at $5.2 trillion.
On this reading, says Schroeder, Artifical General Intelligence is frightening because it has the potential to accelerate all of this:
From this perspective, learning to govern AGI means learning to govern the rich and powerful in general. … It’s not about what AGI might want; that’s a red herring. It’s about what it is, as an enactive being.
I found myself recently in a disagreement with someone who was telling me that Artificial Intelligence was an existential risk to mankind. Until recently I used to have an open mind on this—might be, might not be. All the same, in this conversation, I mentioned that if I was worrying about existential threats, AI came quite a long way down the list, partly because there are multiple more immediate risks.
In the queue ahead of it, climate change, certainly; biodiversity collapse, certainly; nuclear weapons risk, absolutely certainly; probably, also, the large inequalities of wealth on the planet that are capable of causing other systemic risks. Avian flu keeps me awake, along with other pathogens. Even in the technology space I would look at our dependence on space-based PNT to run many critical technologies, and an increasingly crowded lower earth orbit, before I started worrying about AI.
I could go on: I’ve written here over the past few months about the lack of business models for AI, the cost of continuing to developLarge Language Models (the improvement in performance is linear, the increase in cost is exponential), and the environmental issues about consumption of power and water.
Right on cue, Amazon has been shifting data centre work out of Ireland, it seems, because the Irish energy grid can’t take the demand. And the technology academic and columnist John Naughton has a piece in The Guardian this weekend suggesting that we’re just about to hit AI’s peak hype moment.
What my conversation made me realise is that almost all of the discourse that argues that AI is an existential risk is being driven by white men of a certain age who live in affluent parts of the rich world.1 Actually, they are not always the sort of people who usually worry about imbalances of power, at least not from a democratic perspective.
In contrast, many of those who are critical of actually existing AI seem to be women or people of colour. I don’t know what that means yet—I need to reflect on it. But: I am fairly sure that if AI were an authentic existential risk, the social base that was speaking to it as a concern would be a lot more diverse. As it is for climate change, biodiversity risk, and nuclear disarmament, for example.
And in the meantime, like privileged white men do, they’re clogging up the channels with their noise.
2: Turning methane into an asset
I listened to a talk last week on the methane problem on farms. It was given by Neil Ward of UEA and Andy Atkins, Chief Scientists of the IFEAA, and organised, online, by the AFN Network and the IFEAA. The AFN promotes research to accelerate net zero within the food system; the IFEAA works to reduce “fugitive emissions”, of which more late. (Disclosure: I did some work for the AFN Network last year).
The numbers on emissions from the food system are generally well-known. The agri-foods sector accounts for 23% of UK emissions, and within this agriculture accounts for 10-11% of UK emissions. Livestock represents ‘upwards of 60% of agriculture emissions’, and the majority of these come from cows, with some from sheep.
(Holstein-Friesian dairy cow. Photo: www.ars.usda.gov. Public domain.)
So worrying about the emissions from livestock farming is not a trivial concern—circa 6% is about twice as much as aviation, or emissions from data centres, for example.
These are also methane emissions. Although they do not last as long as in the environment—about ten years, compared to hundreds of years for carbon—while they are active they trap far more heat, perhaps 100 times as much. Methane is also a source of air pollution.
This means that anything we do to remove methane from the atmosphere, as climate change ticks inexorably towards 1.5 degrees of warming, is a good thing. The UK is also a signature to the Global Methane Pledge, under which it has committed to reduce methane emissions across a range of sectors, including agriculture. The House of Lords Environment and Climate Change is collecting evidence on progress on this at the moment.
Emissions from slurry
The historic view has been that most of this comes straight out of cows (‘mostly burps’, said Neil Ward) but recent research suggests that this is because the emissions from slurry has been under-counted. Slurry is what you get when you scoop up cow manure and stick it in a tank or a hole in the ground until you get round to spreading it on the fields in the spring.
The talk reviewed quite a lot evidence that suggested that the IPCC data under-estimated methane released from slurry pits by as much as a factor of three, because it used top down assessments that hadn’t been updated over the last couple of decades.
One result of this has been that because the impact of slurry pits was thought less important, there has been less research into the issue. But bottom up assessments, in Canada and in the UK, consistently increases the estimates of how much methane is released from slurry pits.
The good news, though, is that this is both a soluble problem, and even better, a problem that if solved in the right way produces energy and economic benefits for farmers.
It’s a soluble problem because once the slurry is in the pit, it’s a concentrated source, so you can manage the methane basically by putting an airtight lid on the pit. (Not any old airtight lid: I’ll come back to that.)
And because methane emissions have a much bigger impact on warming than CO2 emissions, anything you do has an immediate effect. I think I heard Andy Atkins say that if we could capture “fugitive” methane emissions from farms it would push back climate warming effects by a year, but this seems like a huge effect so I’m not sure I trust my notes here.
Energy source
The good news here is that if captured methane becomes an energy source. Andy Atkins talked a bit about this—more in the Q&A than in the talk:
If we can capture these emissions at source, there would be economic, social and environmental benefits, going into poorer rural communities.
He summarised the benefits in this way:
Farms tend to be at the remote end of the grid, and sometimes have energy issues because of this;
Methane is an energy vector that when released damages both the environment and health (because of worsened air quality);
Capturing it on the farm rather than letting it escape generates either energy that can be used in place of fossil fuels or can be sold to generate funds for the farm;
At the moment we’re still researching the scale and the benefit, because slurry sites are hard to monitor;
But at the very least it gives farmers more control over their costs and changes the economics of livestock farming;
All in all, given that rural communities tend to be poorer, it is an effective form of levelling up.
One of the leaders in this area is the Cornish company Bennamann, which I went to have a look at online after the talk.
The way the technical model works is that the slurry collection and storage system minimises the loss of gas, and the methane is then extracted and processed. It needs a ‘gas tight’ cover, rather than just one that stops rain getting in. The process works with any farm that has animal waste.
(Source: Bennamann)
Bennamann is retrofitting slurry sites in some of its pilots. This requires a different type of technology from that used in the oil and gas industry to capture fugitive methane emissions. The emissions from slurry are diffuse, slow moving, and under low pressure.
Bennamann has a charming video showing otherwise fugitive methane molecules being rounded up by its technology.
Bennamann is now a part of CNH Industrial, a global business that provides services principally to the agriculture and construction sectors.
A news report on one of Bennamann’s pilot farms in Cornwall reports that
(Bennamann Technical Director Chris) Mann said his company's process reduces the carbon footprint of Hoare's farm by around 75 percent. Currently, the farm is producing more gas than it needs, so the local county is buying some of the carbon-friendly fuel to help the community... The next step is developing a generator to take the farm completely off the grid.
I’d like to see more of the financial numbers on all of this, and get an idea of how they might benefit from economies of scope and scale. All the same, the long term consolidation of the dairy industry means that dairy farms are now big enough to make the investment pay. (The average herd size of a dairy farm has increased from 21 cows to 150 cows over the past fifty years).
Neil Ward also pointed out that although there were complex wider arguments about the benefits of regenerative farming as against intensive farming, and this may eventually lead to different approaches to livestock farming, this approach works within the current intensive model.
The talk brought together the interested community effectively—there was a lot of expertise on display in the chat. If you’re interested in this area, there’s a “deep dive” follow up event in person in Bristol on 16th May, with online access as well.
j2t#561
If you are enjoying Just Two Things, please do send it on to a friend or colleague.
For the avoidance of doubt: yes, I realise that there are many people in this demographic who are not AI Doomers. But I haven’t noticed many AI Doomers who are not in this demographic.
Absolutely agree on the over hype of AI as an existential threat, Andrew.