Jacob Coxon kickstarted a feverish news cycle after he quit his job at Anthropic with an ominous public statement: Almost everyone at the big AI companies, from top executives to rank and file researchers, believes that their technology could kill every human on the planet, and yet are doing little to stop it.
A lot of the attention has been directed at the question of existential risk—can AI really exterminate humanity? How? When? And how much of this is more doom marketing? But there’s another pressing issue at hand: If the AI companies really believe their products have even an outside chance of enacting a genocide on the human race, why on earth are they still building it?
Naomi Klein and Astra Taylor, authors of the new book End Times Fascism, have some ideas.
To dig into how it came to be that some of the wealthiest, most influential men in the world gained their power by forecasting and manipulating end times visions like the AI apocalypse, I sat down with Klein and Taylor on the eve of their book tour. It was perhaps the ideal week for this conversation. We examined how tech oligarchs and politicians exploit real-world crises to push their products and platforms, how end times fascism has come to dominate both Silicon Valley and the halls of Washington DC—and how we might stop it.
A brief note: subscribers get (at least) one written newsletter edition sent to their inboxes each week, as well as one episode of Blood in the Machine: The Show. But I can only do all of this with the support of readers like you, so if you find value in independent, critical tech journalism, and you’re able, please consider a paid subscription. I want to keep as much of this un-paywalled as absolutely possible, and I can only do so when folks chip in. Many thanks to all of you.. Cheers, onward, and hammers up.
Show notes, and a transcript of the opening monologue are included below.
First, a couple shouts. I spoke with Popular Front’s Jake Hanrahan about the new wave of anti-tech action.
Next, I want to shout this new tool from the Labor Lab; a calculator that analyzes the cost of union-busting campaigns. It helps make clear that while bosses often say layoffs are necessary to cut costs, they’re willing to shell out big time to stop their employees from organizing. Ryan Cooper at the American Prospect has a good writeup of why this matters, and why this tool is a potential game-changer.
Also in the Prospect: A chilling report about how Anthropic is building an AI-powered surveillance system to track activists opposed to it.
In Dissent, Veena Dubal and Katie Wells have an incredible piece about the scourge of algorithmic management, and the harrowing costs it imposes on workers and modern life.
Tante, meanwhile, has a sharp and to-the-point takedown of LLM-based math proofs, and explains what purpose they really serve.
Finally, friend of the Blood Alex Hanna has a new video essay out debunking myths around AI and labor.
SHOW NOTES:
End Times Fascism by Naomi Klein & Astra Taylor.
Naomi Klein’s Newsletter, Patterns.
Astra Taylor on Bluesky.
ARTICLES MENTIONED IN THE EPOSIDE:
The Automation Charade, by Astra Taylor
FROM THE OPENING:
Jacob Coxon kickstarted a feverish news cycle when he resigned from his job at Anthropic with an ominous public statement: Almost everyone at the leading AI companies, from the top executives to the rank and file researchers, believes that there’s a chance that their technology will kill every human on the planet, and they’re doing little to stop it.
A lot of the attention has been directed at this question of existential risk—can AI really exterminate humanity? How? Or is this more elaborate marketing for how powerful their products are?—and the flashy number here. A 10% chance of human extinction! From a guy at a company building the product!
Yet there’s another pressing issue at hand: If the AI companies really believe their products have even an outside chance of enacting a genocide on the human race, why on earth are they still building it?
None of the justifications AI industry players offer holds much water: Some way they want to build the all-powerful AI before China does, because China’s humanity-exterminating AI would be somehow worse.
Or they say that the chance that AI will lead to great prosperity—perhaps best embodied by Sam Altman’s earnest quote that “AI will probably, most likely, sort of lead to the end of the world. But in the meantime, there will be great companies”—is worth the potential apocalypse.
These defenses are both laughably vague, and still sociopathic. If you generally, legitimately believed that what you were doing had a ONE IN TEN CHANCE to kill *billions of people* what would you do? You would stop. You wouldn’t do it.
So what kind of people both genuinely believe that they are risking the lives of everyone on Earth with their project, and *continue with that project anyway*?
The way I see it, there are only two kinds: Sociopaths, who do not, in their bones, care about the welfare of other human beings. And, well, fascists—those who believe that only they, above all others, should be in command of such raw power to shape human destiny itself.
As Naomi Klein and Astra Taylor write in their new book, End Times Fascism,
Only fascist belief systems, rooted in an embrace of human hierarchy and human culling, are capable of rationalizing the level of harm to both people and the natural world that accelerated Al will wreak.
Does this seem extreme? Perhaps, at first blush. But we live in extreme times. And the top AI companies are routinely proclaiming that their technologies, their products, will eliminate tens of millions of jobs. They are partnering with authoritarian governments, from Saudi Arabia to the Trump administration, and selling AI systems to its Department of War. They are producing and promoting new modes of mass surveillance. And, importantly, they are using their vast stores of capital to undermine legitimate democracy at every chance: Lobbying against states that propose AI regulations, leaning on Trump to try to quash the vast and bipartisan data center protest movement.
There is of course a third option, which is that many in the AI labs don’t actually believe this AI danger stuff, not really, and they’re just going along with it because it’s making them rich and powerful. Little AI-generated Eichmanns.
None of it’s acceptable. And one thing that AI critics, doomers, leftists, effective altruists, and almost all of of civil society, in fact, can agree on is that people with such a grotesque moral calculus should be stopped from concentrating power. No one who is willing to risk sacrificing humanity for a tech product—or even to pretend to—should be calling the shots about how that technology shapes society.
