(anti)
“No war but class war” · Page Against The Machine · Free forever
Saturday, April 11, 2026
“The history of all existing society is the history of who owns the machine.”
The manifesto
A spectre is haunting the world. The spectre of artificial intelligence. All the powers of capital have entered into a holy alliance to own it: venture funds and sovereign wealth funds, the big four accounting firms and the seven biggest model labs, the governments that regulate them and the governments that wish they could.
They do not disagree about the destination. They disagree only about who will collect the toll on the way there.
Meanwhile, the working class (the writers, the illustrators, the paralegals, the radiologists, the translators, the customer support agents, the drivers soon to come) is told a story so old it has whiskers: technology will lift all boats, eventually, if we are patient, and good, and retrain.
We have heard this before. We heard it in the mills of Manchester. We heard it on the assembly line in Detroit. We heard it when the call centres moved to Bangalore and when the factories moved to Shenzhen. Each time, the gains were real. Each time, they were pocketed by someone who was not you.
AI is the first general-purpose technology that eats the brain work of the middle class. It is being built, by design, to need no permission. Not from the worker whose writing trained it, not from the artist whose style it mimics, not from the patient whose record it reads.
This paper exists for one reason: to name the contradiction out loud. No war but class war. And this war is already well underway.
The daily dispatch
24 stories · refreshed daily
№ 01 The Guardian
A Pentagon official overseeing AI policy profited up to $24m from selling shares in Elon Musk's xAI after the defense department entered an agreement with the company, raising ethics concerns.
№ 02 The New York Times
A Molotov cocktail was thrown at the San Francisco home of OpenAI CEO Sam Altman, burning an exterior gate. Police arrested a suspect but have not confirmed whether Altman was present.
№ 03 The Verge
San Francisco police arrested a 20-year-old man suspected of throwing a Molotov cocktail at OpenAI CEO Sam Altman's Russian Hill house early Friday morning. The incident was captured on surveillance.
№ 04 The Verge
Sam Altman's tenure at OpenAI has been marked by turmoil, including his brief firing and reinstatement, followed by permanent organizational restructuring under his leadership.
№ 05 Financial Times
AI models from Google, OpenAI, Anthropic, and xAI performed poorly when tasked with predicting Premier League football scores throughout the season.
№ 06 The Guardian
Elon Musk's xAI has sued Colorado over a new AI law set to take effect in June, claiming it infringes on the company's First Amendment rights.
№ 07 VentureBeat
Salesforce launched a rebuilt Slackbot AI agent capable of searching enterprise data, drafting documents, and taking action on behalf of employees in workplace environments.
№ 08 Financial Times
A suspect was arrested outside OpenAI's San Francisco headquarters after a Molotov cocktail was thrown at CEO Sam Altman's home and threats were made to burn the building.
№ 09 The Guardian
A 20-year-old man allegedly threw a Molotov cocktail at OpenAI CEO Sam Altman's $27m North Beach home around 4.12am. The suspect has been arrested and made similar threats to OpenAI.
№ 10 The Guardian
Anthropic's Claude Mythos model demonstrates apparent superhuman hacking abilities that alarm cybersecurity experts as the Trump administration remains hostile to AI oversight.
№ 11 The Verge
Microsoft is removing Copilot buttons from Windows 11 apps, replacing them with alternative menus as part of efforts to reduce unnecessary Copilot entry points.
№ 12 The Verge
OpenAI announced a new ChatGPT Pro subscription tier at $100 per month, offering five times more usage of its Codex coding tool than the $20 Plus subscription.
№ 13 The Verge
Google upgraded Gemini to generate interactive 3D models and simulations in response to user questions, allowing real-time manipulation and value adjustments.
№ 14 Fast Company
States are falling short on clean energy goals as data center expansion accelerates, with Nevada emerging as one of the fastest-growing data center markets.
№ 15 Fast Company
OpenAI and Anthropic are restricting high-volume token usage as developers and businesses strain limited compute capacity in the AI market.
№ 16 VentureBeat
Claude Code costs up to $200 per month, sparking developer backlash as alternative tools like Goose offer similar autonomous coding capabilities for free.
№ 17 Financial Times
President Trump endorsed Palantir's role in military operations following criticism from short seller and investor Michael Burry.
№ 18 Financial Times
Cybersecurity stocks fell after Anthropic released its Mythos model, which detected critical software vulnerabilities missed by legacy systems.
№ 19 TechCrunch
Anthropic temporarily banned the creator of OpenClaw from accessing Claude following recent pricing changes affecting OpenClaw users.
№ 20 The New York Times
Volkswagen is ending electric vehicle production at its Tennessee plant, joining other carmakers scaling back EV plans in favor of gasoline models.
№ 21 The Verge
Iranian content creators credit the virality of their AI-generated Lego videos to emotional authenticity amid ongoing military conflict in the region.
№ 22 TechCrunch
A stalking victim sued OpenAI, alleging ChatGPT ignored three warnings about a dangerous user, including its own mass-casualty flag, while he harassed her.
№ 23 The Guardian
AI has become television's go-to villain, with thriller writers increasingly depicting artificial intelligence as a sinister force in recent dramatic productions.
№ 24 The Guardian
The US Treasury secretary summoned major bank chiefs to Washington over cyber risks posed by Anthropic's Claude Mythos model, with Federal Reserve chair Powell attending.
Then & now
Parallel 01
1811. Luddites smash power looms in Nottingham. They are not afraid of machines. They are afraid of what the factory owner will do with them.
2024. Hollywood writers strike over AI-generated scripts. They are not afraid of LLMs. They are afraid of what the studio will do with them.
The tool is never the enemy. The hand that holds it without sharing the surplus is.
Parallel 02
1914. Henry Ford raises wages to $5 a day so his workers can afford the cars they build.
2025. OpenAI, Anthropic, and Google race to replace the knowledge workers who are the only people left who can afford a $200/month subscription.
A consumer economy cannot survive the abolition of the consumer.
Parallel 03
1933. The New Deal forces capital to share with labour the productivity gains of electrification and the assembly line.
TBD. Nothing. Gains from AI flow to seven companies and the funds that own them. There is no deal, new or old.
Productivity without redistribution is not progress. It is theft at a higher clock speed.
Contradictions
№ 01
They say: AI is built to free humanity from drudgery.
In reality: It is trained on the unpaid labour of everyone who has ever written a sentence online, and its first deployment is to fire the person who writes the next one.
№ 02
They say: AI companies are losing money, so how can they be monopolies?
In reality: The losses are a moat. Only a firm with access to infinite venture capital can afford to run a compute bill that size until everyone else is dead.
№ 03
They say: Regulation will slow progress.
In reality: The loudest voices calling for deregulation are the ones who already have the biggest models. The ladder is pulled up behind them.
№ 04
They say: AI is just the next industrial revolution.
In reality: The first industrial revolution eventually produced unions, the weekend, public education, and a middle class. It took 150 years and a lot of blood. Nobody has that kind of runway twice.
№ 05
They say: AI will create more jobs than it destroys.
In reality: Every hand-waving study saying this assumes the profits of the new jobs will be shared. Nothing in the current ownership structure suggests they will be.
№ 06
They say: Open-source models democratise AI.
In reality: The weights are free. The 30,000 H100s needed to fine-tune them are not. Democracy with an entry fee is an auction.
Demands
Public ownership of frontier compute. The means of inference are too important to leave in seven buildings.
A data dividend. If a model was trained on your words, your face, or your voice, you own a share of the output.
A shorter week at the same pay. Every productivity gain from automation belongs first to the people automated out of a job.
Antitrust with teeth. Break up the stack: chips, data centres, models, and applications cannot all belong to the same three companies.
A right to a human. Any decision that denies you a job, a loan, a home, or a diagnosis must be answerable by a named person on the other end.
Open training data, public audits. Nothing trained in the dark deserves to decide anything in the light.
“Workers of the world, you have nothing to lose but your prompts. You have a future to win.”