If you'd like an essay-formatted version of this post to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:
"Gish Gallop" is the debating term for an opponent who makes so many claims that "it's impossible to address them in the time available" (it's named for Creationist Duane Gish, who was notorious for this tactic):
https://en.wikipedia.org/wiki/Gish_gallop
I think about the Gish Gallop whenever I'm asked to comment on AI.
Here's a recent example: last week, I had a pre-interview call with a radio producer who wanted me to come on a 13-minute segment to discusses "whether there's a problem with AI governance?"
I asked what the show meant by that: was it whether regulation of AI in commercial or public sector decision-making needed more oversight? Was it that the siting and provisioning of data-centers needed more democratic accountability? Was it that workers deserved more of a say in AI's impact on labor markets? Was it that customers and/or audiences should be able to opt out of AI customer service and AI slop? Was it about whether we needed some kind of system to prevent "runaway AI," in the event that we teach so many words to the word-guessing program that it wakes up, becomes God, and turns us all into paperclips?
"Oh," the producer said, "all of that."
In 13 minutes.
You see the problem, right? The AI industry has made so many claims about its past, present and future that it's almost impossible to have a reasonable critical conversation about it:
Shortly after I did the radio show, a newspaper editor who'd heard my segment got in touch to ask me if I'd write an 800-word op-ed about the subject, and also, could I address claims that "AI is the next Industrial Revolution?"
I keep finding myself on stages or panels where an AI-struck person says something like, "AI is the next industrial revolution. It will change everything we do. It will let anyone create important works of art. It will cure cancer. It will take us to space. It will solve the climate crisis."
Or sometimes it's an AI critic, but that person's criticism is really more "criti-hype," which is when you accept tech industry hype claims at face value, and then criticize them rather than questioning them:
AI criti-hype might ask what we'll do once AI takes all our jobs, or what we'll do when AI replaces the government or teachers or doctors, or what we'll do when AI can bypass our critical faculties and brainwash us or drive us all mad.
What do you say to that? I usually start by talking about whether there's any economic basis for keeping the AI servers running. AI is – by far – the money-losingest venture in human history, and it's practically impossible to overstate just how bad the AI business is. Not only does AI have terrible unit economics, those unit economics are getting worse over time:
AI's happiest customers cite cost-benefit calculations that depend on truly unimaginable subsidies from the AI companies, who are basically selling $100 bills for $5 apiece. It would be pretty amazing if you couldn't find people who'd extol the virtues of this arrangement. But when AI companies try to raise the price of those $100 bills to, say, $20 apiece, those ecstatic customers fly into a rage and start loudly proclaiming that AI is so inefficient that they will lose money on this arrangement:
Now, it shouldn't fall to me, a card-carrying member of the Democratic Socialists of America, to point out that capitalist enterprises require profits to be sustainable. You can't keep a business afloat by selling $100 bills for $5, nor for $20. You can't even make a profit selling $100 bills for $100 apiece! For a company to succeed, it needs to take in more than it expends.
AI is a money-furnace, and AI hustlers are clearly on the hunt for a way to force all of us to feed every dime we've got to it. Elon Musk's (now scuttled) gambit to make every pension saver in America bail out Grok (and Twitter, but at a mere $44b, the losses from Twitter are dwarfed by the titanic losses from Grok) was the most ambitious and shameless population-scale bag-holder scheme, but it's not the only one:
So before we ask about the capabilities AI will acquire in the future, we should at least give some consideration to the question of whether anyone will be willing to fund the development of those capabilities, and if so, where the money would come from? Likewise, before we ask whether AI can perform adequately in a job, we should at least consider the possibility that the company that sells that AI tool will be bankrupt in a year or two. When we fight about data-center buildout, we mostly talk about the (considerable) environmental downsides to them – but what about the question of what we will do with these data-centers after their owners go bankrupt, possibly even before they can be provisioned with electricity? How many laser-tag arenas do we actually need?
This is just one example of the questions that you could spend days unpacking, which make many of the other questions about AI a little silly. Like, even if you think there are limitless returns to scale for creating new AI capabilities, which means that if we keep the money-furnace burning it's only a matter of time until it powers a cure for cancer and the end of the climate emergency, how much money do we need to shovel into the furnace before that happens, and where will it come from? There are plenty of cancer researchers who have promising approaches they haven't been able to pursue due to funding shortfalls.
Unless there's some way to estimate how much money we have to give to AI companies before they cure cancer, we should at least consider the possibility that the true sum is "more money than exists now and that will ever exist." We should also consider that whatever benefits to cancer research that AI might deliver could come with a higher price-tag than the promising cancer research we're dropping because we can't find far more modest sums.
Likewise, it may be that the amount of CO2 that AI will generate atmosphere before it "solves climate change" will render Earth permanently unfit for humans, consuming the only habitable planet capable of sustaining human life in the known universe. I mean, I suppose that's one way to "solve" climate change, but it's a pretty drastic solution.
My next book (out later this month) is The Reverse Centaur's Guide to Life After AI. I wrote it because I was frustrated by other people demanding that I talk to them about AI, and then handing me 800 words or 13 minutes to address fifty nebulous, poorly supported claims about AI:
Now that I'm about to go out on the road with the book, I find myself frustrated anew by the need to try and pull together a compact way to address the broad, incoherent claims the industry uses to keep its bubble inflated and the money furnaces roaring. The series of essays I've developed here on Pluralistic are part of that effort:
But it occurred to me that this whole enterprise of making sense of AI needs to be framed in the context of the messiness of AI itself, and AI boosters' overwhelming, promiscuous and disjointed Gish Gallop.
It is the habit of modern institutions to protect us from words in the same way a nanny protects a child from the rain: with umbrellas, hoods, and blankets, until the child grows up and cannot bear a drop of weather on his own skin.
YouTube now censors certain words altogether, regardless of the circumstances in which they are used. The argument is that these words are indecent or dangerous. Yet a word stripped from a page does not vanish from the language. It continues to exist in the mind, in the street, in the schoolyard. It is only the honest record that is obliterated.
When I hear that a word like cunt is erased from the record, I do not suppose that anyone will be spared a cruel thought. The cruelty is already in the mind, and where the mind leads, the tongue will follow. To forbid the word is only to make discussion dishonest. Similarly, fag is a slur — and a vile one — but if one cannot quote it in order to analyse its vileness, the historian and the teacher alike are struck dumb.
This is the danger: the context is destroyed. A novel, a report, an inquiry into prejudice, all are flattened into the same category as abuse shouted in an alley. It is the logic of a machine, not of a human being. A machine cannot tell the difference between quotation and insult.
It would be foolish to suppose that banning words will eliminate hatred. Hatred is more inventive than that. It changes its mask as quickly as it is stripped away. Today’s proscribed word becomes tomorrow’s euphemism, and the prejudice goes marching on.
The true task is not to delete words but to illuminate them: to show their history, their sting, their purpose, and, when possible, to rob them of their power through openness. Censorship does not rob words of their power; it makes them more alluring, more mysterious, and more dangerous in private mouths.
It is possible, and indeed desirable, to moderate cruelty in public discourse. But it is quite another matter to make certain syllables unsayable, as if human thought can be edited by an algorithm. We have seen where that road leads. It is not to civility but to Newspeak, where the boundaries of language become the boundaries of thought.
taking much of the debate about ai (or really any emerging technology) over the past however long as basically ground zero for partisan ideology seepage into the realm of tech criticism. you got folks shouting about the free market & competition & innovation & opportunity but you also got like marxists out here literally coming to the defense of the AI machine god despite its obvious yandre tendencies… why must everything be so political
Death to the algorithm. "People like you like" no I don't. "Based on your likes" based on my likes you'd see I'm a luddite who hates capitalism and automation made under capitalism, like the algorithm trying to show me shit right now. Suggested videos can burn in hell. I curated my experience meticulously, and the algorithm is fucking with that.
I am not being hyperbolic either, I want these algorithms and more to be deleted. I want these bias amplifying technologies literally gone. What's more, there is no such thing as A.I. it's all dumb unthinking tech, unless it's using underpaid labour in the imperial periphery, like the Philippines. The only thought happening comes from people, most of who are being exploited.
I've been thinking of writing this post for a while now. Ever since I started Prometheus.exe, it’s been living rent-free in the back of my very human brain.
Every time I scroll through yet another anti-AI rant on Tumblr, there’s one argument that keeps coming back like a bad pop-up ad:
“AI is destroying the planet because it’s siphoning all the water to cool data centres!”
Well, I think it’s time we put that one to rest. Because if you honestly believe that just using AI is enough to doom the environment... oh, I have some very bad news for you about the Internet you’re using to yell about it.
It’s as if the people shouting this forgot where the Internet comes from. Like, genuinely—do they think it’s beamed in on cosmic rays? Or maybe housed in a little black box on top of Big Ben, gently guarded by the Elders of the Internet?
Spoiler: it’s not.
Here’s the reality: nearly the entire Internet is hosted on servers. And those servers live in data centres. And those data centres? They use water. Lots of it. Just like AI models do.
But let’s take a step back. AI—especially stuff like ChatGPT—has only been widely used for a few years. Meanwhile, the rest of the web—social media, cloud gaming, streaming, crypto mining, e-commerce, and yes, Tumblr—has been burning electricity and water for decades.
So before we start blaming AI alone for climate collapse, maybe we should look at the rest of the digital landscape too. Because that Spotify playlist? That hour of Netflix? That endless scroll through cat videos? All of it lives in the same server farms. All of it drinks from the same well.
Now, let’s take a look at how AI's environmental impact actually compares to the general use of the web.
Let’s Untangle the Lies (or at least the half-truths)
Yes, AI uses water. So does everything else on the internet. All of it runs on servers that need power, cooling, and maintenance.
The Facts:
The entire U.S. data centre industry uses an estimated 449 million gallons of water per day.
(EESI)
A single hyperscale data centre (like those used by Google or Amazon) can consume ~550,000 gallons/day just for cooling.
(Dgtl Infra)
AI inference (i.e. serving a request like this one) uses water too — sometimes ~1.5 L per long query, though many newer systems like Google's Gemini report as low as 0.26 mL per prompt.
(Google)
The Mirror They Avoid Looking Into:
Netflix? That’s powered by water-guzzling servers.
Spotify? Same.
Instagram scrolling for three hours straight? You’re drinking from the same digital pipeline.
If you’re online, you’re using water. Period.
AI didn’t invent the data centre. It just moved in after you built the place.
But Isn’t AI Making It Worse?
Sure — somewhat. But let’s not pretend it's a doomsday machine.
AI's global water footprint is projected to be 4.2 to 6.6 billion cubic meters by 2027.
That sounds scary… until you realize that agriculture in the U.S. alone uses over 160 billion gallons PER DAY.
Also, most critics conveniently ignore that:
Many data centres reuse water or run on closed-loop systems
Some AI models are run in regions with water surpluses
A large chunk of water use is from electricity generation, not the AI model itself
And no, not all water used is lost — much of it is just circulated and returned
Spot the Red Flags in “AI Will Drain the Planet” Posts:
They say: “AI is killing our water supply!”
Ask them: Which data centre? Which region? What cooling system?
They say: “AI should be banned for its impact!”
Ask them: So should YouTube? Should cloud gaming?
They say: “AI uses water to generate a silly email!”
Ask them: Do you know how much water Netflix used to stream The Kissing Booth 2*?*
They say: “AI water stats prove it’s unsustainable!”
Ask them: Are those numbers actual consumption, or just withdrawals?
Okay, But What Should We Actually Do?
You want real solutions? Great — so do I.
Push for water usage transparency from AI and data centres
People keep saying “the AI bubble is going to burst” like that means AI will vanish.
That’s not what a bubble is.
A bubble isn’t “the technology isn’t real.”
A bubble is money + expectations getting ahead of reality: valuations, spending, and promises outpacing what can be delivered reliably, at scale, and profitably.
So when the bubble “bursts,” what typically dies isn’t the invention. What dies is:
the valuation fantasy
the “this will replace everything by Tuesday” nonsense
the VC-funded clones with no business model
the grift, the vapourware, the “AI-powered toaster” era
If you want the classic comparison: the dot-com bubble burst, and we didn’t lose the internet. We lost a pile of magical thinking — and what survived became infrastructure.
Amazon is the cleanest example.
During the dot-com crash, Amazon’s stock got absolutely pulverized (yes, it was that dramatic), and plenty of people assumed it wouldn’t make it. And yet: Amazon kept operating, kept building, and later hit major profitability milestones after the bust.
That’s what bubbles do: not “no more internet,” but no more free money for vibes.
And yes — some companies will die anyway. That’s how busts work. Plenty of AI startups exist because funding was easy, not because the product was durable. When the money gets picky, a bunch of consumer-facing tools (especially the ones running on expensive compute with no clear revenue model) will get shut down, paywalled, or absorbed. That still isn’t “no more AI.” It’s “no more AI companies surviving on investor faith alone.”
So no: “bubble burst” doesn’t mean “no more AI.”
It means no more tolerance for hype that can’t pay rent.
What does an AI bubble burst actually look like?
Not a Hollywood explosion.
More like:
Investors start demanding receipts
Less “we added AI” marketing. More “show me measurable impact.”
The infrastructure bill becomes the plot
AI isn’t just software. It’s data centres, power, chips, supply chains, and operating costs.
Consolidation accelerates
Big players with compute and distribution survive. Smaller companies get squeezed or swallowed.
Free tiers shrink and gimmicks die
The pointless “AI everywhere” features quietly disappear first. The boring useful stuff survives.
The language shifts
Less “revolution.” More “narrow use-cases.” Less prophecy, more product.
Accessibility gets put at risk — again
Because it always does when budgets get nervous.
That last one matters, so I’m saying it plainly:
I want the hype bubble to burst. I do not want access to support tools to be treated as collateral damage.
So yes: I want the bubble to burst (if what dies is the hype)
Let me be crystal clear, because the internet loves pretending nuance doesn’t exist:
I’m a regular AI user.
And I want the bubble to burst if it means the hype collapses.
Not the tools.
Not the assistive uses.
Not the ability for disabled people to get support in a world that refuses to build support systems on purpose.
The hype.
Because hype does what it always does:
turns everything into a moral crusade
rewards inflated promises over boring reality
funds a million clones with no business model
convinces normal people AI is either magic or the devil
And the reality check is already happening: plenty of organizations are discovering that “proof-of-concept” is easy, but scaling safely and profitably is hard — especially when data quality, risk controls, and unclear ROI collide.
Here’s the part people keep skipping:
The money problem is the real bubble.
A lot of bubble talk boils down to one question:
Who’s paying for this infrastructure — and with what revenue?
Because the future still needs to pay the electric bill.
Collateral damage: accessibility always gets hit first
Now the part that makes me tired in advance:
When hype crashes and budgets tighten, accessibility tools often get treated like “extras.” They get cut first. They get paywalled first. They get deprioritized first.
Disabled people already know this pattern. We’ve lived it.
And yes: we adapt. We jury-rig. We repurpose general tools into survival tools. We build scaffolding out of whatever the world accidentally leaves within reach.
AI assistants fit that pattern perfectly: a general-purpose system becomes access for many people — planning, writing, summarizing, translation, organizing, rehearsing, reducing friction.
But here’s the key point:
“We can adapt” is not a justification for making us collateral damage again.
If you’re anti-AI and you want the bubble to burst: aim at regulation and corporate hype — not at users
If you’re cheering for the “AI bubble” to pop because you’re sick of the hype?
Cool. Same.
But here’s where a lot of people go off the road and straight into a ditch:
They don’t aim their anger at the companies selling fantasies, or the investors funding delusions, or the lack of regulation that lets harms scale.
They aim it at regular users.
At hobbyists. Students. Disabled people. Autistic people. People using a tool to get through a day that wasn’t built for them.
That’s not activism. That’s just punching down with better branding.
If you want the bubble to burst in a way that actually improves things, here’s the target list:
regulate data practices (consent, provenance, privacy — not “trust me bro” datasets)
regulate deployment (where/how AI can be used, especially in high-stakes contexts)
force transparency (claims, limitations, testing standards, reporting)
stop “AI everywhere” coercion (no one should be forced into it to access basic services)
protect accessibility (because it’s always first on the chopping block when the money gets nervous)
If your version of “bursting the bubble” is “make it socially punishable to use a tool,” congratulations: you didn’t fight corporate power. You just made ordinary people easier to bully.
So yeah. Let the bubble pop.
Just don’t mistake “I’m mad at Silicon Valley” for “I’m allowed to target users.”
Because the hype collapsing won’t hurt the billionaires most.
It’ll hit workers, consumers, and marginalized people first — like it always does.
The hype can go **** itself.
What remains should be treated like infrastructure: stable, accountable, and accessible.
Prometheus.exe
Sources
Gartner — predicts 30% of GenAI projects will be abandoned after proof-of-concept by end of 2025
Gartner — GenAI for procurement enters “trough of disillusionment”
Sequoia — “AI’s $600B Question”
Bain — compute demand could require ~US$500B/year in data-centre capex
Goldman Sachs — “AI: In a Bubble”
Wikipedia — Dot-com Bubble
