Lobbygruppen wissen genau, wie sie sich einen Weg in die Politik bahnen können. Dagegen braucht es ein Bewusstsein und institutionelle Schranken. mehr...
"Es braucht institutionelle Schranken für Lobbyaktivitäten, wenn sich das neoliberale Credo des „schlanken“ Staates nicht länger in politischen Entscheidungen niederschlagen soll."
Spend enough time in the world of corporate sustainability, and you’ll hear a reproachful complaint: it’s unfair to criticise companies that are saying the right things on climate. If you let the perfect be the enemy of the good and punish the few trying to make a difference they’ll get spooked and not bother at all.
I don’t really buy that. Faking climate action can often be more harmful than doing nothing. For-profit companies gain financially by lying about the harm they’re doing. But even beyond that, fabricated cleanliness instils complacency and dissuades us from asking deeper systemic questions. The integrity of claiming to be doing good should be protected, and doing that involves critiquing those making the loudest, most strident claims of goodness.
I wrote about Microsoft as a key recent example: a company with rapidly growing emissions but a persistent claim to climate hero status. I found that the harm they do far outweighs the good. And I think even non-profit organisations create permission space for the broader trend of generative digital bloat currently burning through my digital world like a raging bushfire. I want to tell you about a recent example of this problem.
Why do you need to kill your search engine
Ecosia is an alternative sustainability-themed search engine that operates as a not-for-profit. The idea is that for every search you perform, the company plants a tree (it has planted millions). It has been around for a while, and always felt pretty harmless. There is nothing at all to dislike about a search engine that effectively functions as a fundraising tool for tree-planting and clean energy projects, and it seems to have actively funded a pretty stunning number of good, environmentally and socially-aware climate projects.
Their latest change is a step away from that good track record. That is because they’ve followed the big-tech trend by installing a layer of automatic plagiarised text generation on top of search results.
Instead of providing a link to a blog post Ketan has written, Ecosia will call on a software service that has ingested all of my written work, and that service will re-word and re-publish my work , often botching and mutating it thanks to its clumsy probabilistic word-guessing design. This isn’t “search” anymore: it’s a roadblock that stops you searching and funnels you away from engaging with human-created content.
As a person-who-writes-internet-content, you can guess how deep this burns. It isn’t just traditional plagiarism: it is unfeeling, automated, non-consensual and, worst of all, wrong enough to matter but right-sounding enough to be convincing. This isn’t an evolution of search: it is the active murder of search.
To demonstrate this point, I entered a question a normal human could easily answer with traditional methods: which books has the climate writer Ketan Joshi written? Ecosia immediately blended me with another brown guy with the same name:
Even if you use ‘traditional’ search, you still get an AI overview that blends me with travel-writer Ketan Joshi at the top of the page that you can turn off, but will always appear by default. A query I tried last week was worse: it got the title of my book right but manufactured the exact opposite meaning with the subtitle (in addition to attributing a bunch of books I didn’t write to me):
And when you ask it for quotes from the climate analyst Ketan Joshi, it either fabricates sentences and puts them in quotation marks, or pulls real quotes written as comments, or it even attributes lines I was criticising to me:
Quote 1: it's a COMMENT on my @techwontsave.us chat posted on youtube!2: from my IEA AI report thread; not said in the post or the quoted images *at all* 3: A quote from a video FROM A DC DEVELOPER, not from me!! I posted the vid of him!!4: headline, not a quote?5: Also a headline…..
I mention this because it’s the important half of the sustainability question: why do it in the first place? “Everyone else is doing it” is not only not a justification, it should in fact be the thing that triggers more hesitation and scrutiny of what’s being implemented.
I assume the actual reason is the same as most other search engines: users acquire the empty calories false satisfaction of a more-easily-obtained answer. By definition, people searching for an answer don’t know the answer. That means they don’t detect how egregious and false the result is (particularly when the text engine is tuned to sound plausible and authoritative, and never output “I don’t know”).
It’s worth mentioning here that iFixit, an organisation that has been fighting against device waste through advocating for repair rights, also recently implemented a “chatbot” function on top of their many user-created repair guides. The verdict: “Having tried it, I would definitely not trust iFixit’s FixBot to guide amateurs like me through a pricey or dangerous repair”. On top of that: they too do not disclose any information about energy use compared to the old way of doing things. What is the point?
Although Ecosia is a not-for-profit, they seem to be re-enacting the for-profit erosion of human knowledge and the active murder of the open web currently being orchestrated by Google and the other tech giants. But, it’s green, right?
Green slop > Grey slop
As an extension of their green image, they’ve also presented this text-generating tool as the “world’s greenest AI“:
As I bet you already know, the mass-generation of text, images and videos using machine learning tools trained on repositories of all human-created digital information consume a lot of energy, and in doing so, materially boosts the burning of fossil fuels, doing damage to our atmospheric safety.
Ecosia isn’t doing this for profit, but the underlying idea of their claim – that “AI” can be green – is important for the profit-seeking tech industry, striking at a fast-growing tension between the goals of companies and the biophysical vulnerabilities of living things.
So, let’s check some of their claims.
"Reducing AI’s footprint isn’t enough — we’re here to make a positive impact. That’s why we generate more renewable energy than our AI features use, from 100% clean sources like solar and wind"
You need two numbers to check this: how much renewable energy they generate, and how much energy their “AI” features use. I asked Ecosia CEO Christian Kroll about both, and he referred me to this post, which very much does not contain either.
Honestly, this claim doesn’t even sound that implausible, which makes it even more odd that the numbers aren’t being shared. Sharing these types of figures openly would actually be a genuinely useful exercise, and give us an open, transparent insight into their energy consumption both pre and post text-generation. Despite that: nothing.
More problematic is the idea that Ecosia is “carbon negative“1 thanks to both causing emissions but also funding renewable energy. This is the logic of carbon offsetting and it fundamentally doesn’t make real, physical sense: no matter how much more renewable energy you fund, emissions are emissions and they still cause climate change. We solve this problem when we stop emitting.
"We use OpenAI’s GPT-4.1 mini for the best balance of performance and efficiency. It uses far less energy than larger models, while still delivering the answers you’re looking for"
Not entirely sure this needs saying but OpenAI is comfortably an industry leader in both lacking transparency and actively, directly incentivising the burning of fossil fuels to power their energy-hungry data centres. I cannot think of a worse choice, when it comes to trying to make a chatbot “green”.
Close to 1 full goddamn gigawatt of gas (open-cycle, so the most inefficient kind) for the "Project Jupiter" site in New Mexico.This whole article is a stunning illustration of how data centres are incentivising new fossil infrastructure: eastdaley.com/daley-note/p…
Notably, they also seem to exclude the training costs, or any analysis of which servers are doing the training or results (“inference”) (and what grids they’re operating on).
"We use tools like the AI Energy Score and Ecologits to select efficient models and track their energy use — keeping our process transparent, and ourselves accountable"
"As a not-for-profit company, we can afford to do things differently. AI Search uses smaller, more efficient models"
As I mentioned, Ecosia don’t disclose energy use: not for a single query, and not for their entire organisation (particularly before and after the implementation of incorrect text generation replacing links to the open web). It would be exceedingly easy for them to implement a live energy usage and estimated emissions score for interactions with text generation and compare them to estimates of traditional search. A not-for-profit is better-placed than most to do something like that. But it really seems like they want the accolades of “green AI” without having to disclose a single actual number.
"Prefer the classic experience? You can turn Overviews off with a single click"
"We avoid energy-heavy features like video generation altogether"
9 months ago, Ecosia posted a clip of prominent AI and sustainability expert Dr Sasha Luccioni highlighting that one of the reasons she uses Ecosia is thanks to them not forcing the use of generative systems on search users. Unlike Google and Microsoft, Ecosia does allow users to ‘turn off’ generative summaries.
But I think this is important: while they do offer the option to ‘turn off’ the activation of text generation, it is on by default. The site’s design uses ‘dark patterns’ to usher you into replacing web search with inaccurate text plagiarism by default – hitting ‘enter’ takes you to a search, but the only button on the bar is to activate the chatbot, resulting in a confusing interface at best:
As I wrote in my post about Google, there has been a very noticeable trend towards actively pressuring and mandating chatbot use. Ecosia seem to sit in the middle: even offering an option to turn it off is radical, but having it on and prominent by default will do plenty of harm that could be avoided through much more honest communication.
Hedonistic sustainability
A few months ago, I visited my brother in Denmark. He works in an oddly tall office building in Copenhagen, and from it, you can see the gorgeous near-shore collectively-owned wind turbines. Directly in the foreground, obscuring the view of at least one of them, was this thing:
That is a giant facility that generates electricity by burning household waste.
When you burn a plant that sucked carbon from the air as it grew you’re hypothetically just returning that carbon to the atmosphere: ‘carbon neutral’. But when you burn the plastic packaging it came in, that plastic is made from carbon extracted from deep underground, which you’re then transferring to the sky. That is carbon pollution, and ‘waste to energy’ (WTE) plants burning plastic are becoming a shocking contributor to rising global temperatures.
“Copenhill” was meant to be different. It was meant to puff out its smoke in ‘smoke rings’ that would remind Copenhagen residents to reduce waste (“those released at night will be illuminated by lasers connected to a heat-sensitive tracking system”). That never happened, and importantly, neither did the planned carbon capture and storage project. That failure in particular directly contributed to Copenhagen missing its 2025 net zero pledges, despite the city making massive progress thanks to wind power, reduced consumption and electrification.
The rate at which locals recycled plastic instead of discarding it was underestimated: meaning the oversized plant now has to import waste from other regions to burn to make financial sense. That imported waste has a higher share of plastic and therefore higher emissions2.
You can still ski down Copenhill on a plastic surface. When I visited, the elevator still told me of the impending CCS facility. The bar at the top is sponsored by Coca-Cola, one of the world’s worst sources of plastic pollution.
Copenhill was developed by Bjarke Ingels, of architecture firm BIG. Ingels coined the phrase “hedonistic sustainability”, “to demonstrate how the ‘seemingly contradictory’ ideas of sustainable development and the pursuit of pleasure can, and indeed should, co-exist”. In one 2016 video, Ingels specifically cites the Tesla Model S as an example of where the clean alternative can be better, thanks specifically to its greater acceleration.
"It would be great for the owner of the power plant because he wouldn't have to or she wouldn't have to make expensive ad campaigns or print pamphlets where people could read that her technology was clean because when you go and see it it's like 'wow what is this? This is like a completely different kind of power plant'"
Honestly: this isn’t ‘hedonistic sustainability’. It’s not even really hedonism: the ski slope is pretty mid. I didn’t get the impression from my conversations with locals that it’s much loved or heavily-used feature of Copenhagen’s cityscape. The guy working in the bar at the top looked beyond bored.
Being able to hop on a freely available share e-bike outside my brother’s office and cycle on absurdly safe roads en-route to Copenhill? That really was ‘pure pleasure’. The city-choking and pedestrian-killing over-acceleration of heavy Tesla EVs gets cited as the prime case of ‘hedonistic sustainability’ because I think these folks don’t have a good idea of what either word really means.
Ecosia is embarking on a similar process: offering up a vision of unethical, bloated, oversized and financially unstable overconsumption as compatible with sustainability. If they had simply said “we’re not replacing search with an interface that spits out inaccurate estimations of human content because it’s not worth the climate cost”, and instead invested more time and money into making search as useful and effective as it used to be, it would’ve been genuinely worth celebrating. And outside the capitalist corporate sustainability world, there is a huge, hungry audience for acts of resistance against the corrosive, life-worsening trends being enforced by big tech without any of us asking for it.
You can’t be ‘sustainable’ without asking whether something even needs to be done. This is true for replacing search gateways to the human-created internet with energy-hogging content plagiarism roadblocks. It’s true for massive waste burning facilities, or oversized road tank electric vehicles. These things are not really “hedonistic” or “pleasurable”. They’re anti-social, soul-crushing and demoralising, and they can only be presented as “green” through the lack of any real disclosures.
Ecosia also claims that their actions are ‘removing’ carbon from the air, which is not the case. Their impact is an addition of greenhouse gases into the atmosphere, no matter how much renewable projects they fund. You cannot undo the damage of pollution. ︎
"Although Ecosia is a not-for-profit, they seem to be re-enacting the for-profit erosion of human knowledge and the active murder of the open web currently being orchestrated by Google and the other tech giants."
I’ve written far more than I should have had to on the performative inclusion of “stakeholders” in post-2016 tech policymaking. This meant being invited into meetings with government and decision-makers, being told to “assume positive intent” as all manipulative types like to insist, only to find that your presence was strictly performative. You were either there so that they could tick the box of saying they had engaged with you, before proceeding to do what they were going to do anyway, or you were there so that they could spin your presence as an endorsement of what they were going to do anyway.
Turns out it wasn’t just me – the behaviour was so widespread that some academics have now done a study into how UKGov wields “stakeholder” engagement.
They conclude:
These findings show that the use of stakeholder tends to performatively entrench the existing power of “industry stakeholders” or nameless but clearly already engaged and empowered “key stakeholders”. Meanwhile, they also construct a false sense of inclusion through the non-performative use of generic or “other stakeholders”. This creates significant risk of a veil of accountability, and raises significant questions over established processes such as consultation. When it is unclear who is influencing policy, whose voices and interests are being represented, then the indicators from specific uses suggest that the stakeholder becomes a foil for amplifying historical power and privilege, often on political and/or economic lines, and in doing so excludes the needs of those most affected by technologies who already suffer a lack of agency in how data, AI, platforms and other areas are used to shape their lives.
"These findings show that the use of stakeholder tends to performatively entrench the existing power of “industry stakeholders” or nameless but clearly already engaged and empowered “key stakeholders”"
Das Programm für die Rack & Stack steht in großen Teilen, wir haben Zusagen von einigen tollen Speakern! Bis Ende Januar können noch Vorschläge eingereicht und vergünstigte Tickets gekauft werden. (Golem.de, Software)
I am not a talented person, I’ve never been called “gifted” or anything like that. Anything I can do, anything I achieved took a lot of work and stubbornness to achieve. Don’t get me wrong. I am not saying that I’m “selfmade” and that my position in society, my access to resources, etc. had nothing to do with it – quite the opposite. As a white heterosexual cis-man in Germany I have started live on easy mode. But I do not come from a wealthy background or one with large networks and access to power. I am not associated to any organization that gives me “respectability” or “relevance”. Just a dude with a website who sometimes writes a few things that luckily people read and that got me some opportunities.
But those opportunities did take a long time to materialize. Like the first 5 to 10 years of me writing anything nobody gave a fuck. Which is probably good, there must be many dumb takes in there. But you learn and grow (hopefully) and today I have a modicum of visibility and a handful of people who read what I write and sometimes even pass it on to others. But it took 15 to 20 years to get there. It was a lot of work.
When we talk about “AI” these days, we usually are not really talking about material systems and their actual properties. We talk about a vision, a narrative. About a hope maybe? The hope that we have machines now to get something from nothing.
In the beginning of 2025 Mikey Shulman (CEO of Suno, one of those websites where you can generate AI muzak) guested in a podcast and – maybe accidentally – framed the conversation on “AI” perfectly in two terms. Shulman was talking about his service and how it was democratizing music and whatnot. Let’s look at a quote from the interview.:
“And so that is first and foremost giving everybody the joys of creating music and this is a huge departure from how it is now. It’s not really enjoyable to make music now“
Did you see the two relevant words? Here’s another quote:
“I think the majority of people don’t enjoy the majority of the time they spend making music.“
Now you might think that this statement is a bit … ridiculous. Most people make music for fun (because very few people can live off of it). They make it because they like it or because they just need to express something and music is their language. Music is a business – sure – but it’s also a Hobby (remember those: Things people do just cause they enjoy them without making money off of them?). But joy/fun is not the word I mean. It’s the distinction of creating and making.
Shulman calls using his slop-machine creating (which brings joy) and what other people do making (which nobody likes). I find this distinction revealing but also in a way counterintuitive: Isn’t it this big cultural norm to take pride in your work? That “hard work” is something dignified? Is that not the whole foundation of the “founder”/”entrepreneur” (I always feel like I need to take a shower after having typed that word) and how important they are narrative?
Creating in this understanding is exactly the idea to get something from nothing. Just think it and it exists. There is no process, no obstacle, no materiality, no challenge or struggle. Just the pure joy of creation. I won’t even get into the religious undertones of that distinction and how using “AI” is framed as godlike even though that would probably be fruitful as well. In this reading the “joy” comes from having the artifact and all that comes from that (like being able to sell it, use it, present it to others in order to gain social recognition, etc.). The joy is maximum unbound productivity. Because that’s what everything we do is for, right? Producing. Making a number go up. Creating means decoupling objects from the process of their making, decoupling them from the resources needed for them to be made, decoupling from the work that others needed to put into the systems enabling the creation. It’s the whole “do not let the real world, the other people in it tough me” thing that defines so much of tech bro logic.
In contrast making is about the process. Sure, there’s also something at the end, and artifact, something you might want to have. But the process itself is always part of it. Making is not just about having a thing it’s about the transformative experience of being in this world, interacting with its objects, properties, with other people in order to bring something into existence that means something to it. And every one of those processes leaves a mark on you. Something you learned, something you’ll remember. That cool trick you found in doing something smarter. Or (as it often was and is in my case) a wound or scar of where you fucked up. Making is part of what allows you to become you, to be you. You are in the process, not just the object at the end and its potential use or sale.
And I think this distinction shows why “AI” and capitalism are so deeply intertwined and why there probably is not really a leftist version of it: Capitalism wants to produce more and more with as little cost as possible to create growth and therefore value for those owning capital. “AI” is that: Don’t put anything in really and get something that might be passable to sell. You cannot get better than that really (from a capitalist viewpoint).
But even when thinking about how there is a leftist case to be made for automation (which I think there is) does that fit? Is the leftist case for automation “a lot of shitty products”? No – the point would be to produce the high quality goods people need in a way that is sustainable while still giving people more time to do things they enjoy. Like making music.
“AI” claims that making things is for suckers. But a leftist case for automation only makes sense if it opens up time and space for people to spend their time on things they enjoy. To participate in processes that enhance their lives, their connections to others.
I think that the focus on creating is just the little capitalist devil sitting on our shoulders telling us to produce more.
I think the most radical act today is to just make something. Especially if you are not good at it or if it’s a bit of a struggle. Draw if you’re not good at it. Play the piano even though you’re not great. Make something just for the fun of being in this world, touching it, being in it. Becoming you. Let this radicalize you a bit.
IBM owns Red Hat which in turn runs Fedora, the popular desktop Linux distribution. Sadly, shit rolls downhill, so we’re starting to see some worrying signs that Fedora is going to be used a means to push “AI”. Case in point, this article in the Fedora Magazine:
Generative AI systems are changing the way people interact with computers. MCP (model context protocol) is a way that enables generate AI systems to run commands and use tools to enable live, conversational interaction with systems. Using the new linux-mcp-server, let’s walk through how you can talk with your Fedora system for understanding your system and getting help troubleshooting it!
This “linux-mcp-server” tool is developed by IBM’s Red Hat, and of course, IBM has a vested interest in further increasing the size of the “AI” bubble. As such, it makes sense from their perspective to start pushing “AI” services and tools all the way down to the Fedora community, ending up with articles like this one. What’s sad is that even in this article, which surely uses the best possible examples, it’s hard to see how any of it could possibly be any faster than doing the example tasks without the “help” of an “AI”.
In the first example, the “AI” is supposed to figure out why the computer is having Wi-Fi connection issues, and while it does figure that out, the solutions it presents are really dumb and utterly wrong. Most notably, even though this is an article about running these tools on a Fedora system, written for Fedora Magazine, the “AI” stubbornly insists on using apt for every solution, which is a basic, stupid mistake that doesn’t exactly instill confidence in any of its other findings being accurate.
The second example involves asking the “AI” to explain how much disk space the system is using, and why. The “prompt” (the human-created “question” the “AI” is supposed to “answer”) is bonkers long – it’s a 117 words long monstrosity, formatted into several individual questions – and the output is so verbose and it takes such a scattershot approach that following-up on everything is going to take a huge amount of time. Within that same time frame, it would’ve been not only much faster, but also much more user-friendly to just open Filelight (installed by default as part of KDE), which creates a nice diagram which instantly shows you what is taking up space, and why.
The third example is about creating an update readiness report for upgrading from Fedora 42 to Fedora 43, and its “prompt” is even longer at 190 words, and writing that up with all those individual questions must’ve taken more time than to just… Do a simple dry-run of a dnf system upgrade which gets you like 90% of the way there. Here, too, the “AI” blurts out so much information, much of which entirely useless, that going through it all takes more time than just manually checking up on a dnf dry run and peaking at your disk space usage.
All this effort to set all of this up, and so much effort to carefully craft complex “prompts”, only to end up with clearly wrong information, and way too much superfluous information that just ends up distracting you from the task you set out to accmplish. Is this really the kind of future of computing we’re supposed to be rooting for? Is this the kind of stuff Fedora’s new “AI” policy is supposed to enable?
If so, I’m afraid the disconnect between Fedora’s leadership and whatever its users actually use Fedora for is far, far wider than I imagined.
"IBM owns Red Hat which in turn runs Fedora, the popular desktop Linux distribution. Sadly, shit rolls downhill, so we’re starting to see some worrying signs that Fedora is going to be used a means to push “AI”."
Log in
︎
