Ten confessions from the AI gold rush
They promised you the future.
They sent you the invoice.
Ten books inside the machinery of the AI boom—and the world picking up the tab.
The future has left the meeting. The bill is still here.
The complete collection / 01—10
Every volume opens a different door into the AI economy. Start at the beginning, or enter wherever the bill feels familiar.
How AI Ate the Internet and Called It Fair Use
Read 5 chapters free
How AI Drinks Your Lake and Sends You the Water Bill
Read 5 chapters free
How VCs Invest in Their Own Customers and Call It Growth
Read 5 chapters free
How AI Fakes It Till the Stock Options Vest
Read 5 chapters free
The Invisible Workers Who Teach AI What Not to Say
Read 5 chapters free
How AI Companies Lobbied for the Regulations That Kill Their Competitors
Read 5 chapters free
How AI Watches You While You Watch It
Read 5 chapters free
How AI Closed the Door Behind It
Read 5 chapters free
How AI Lies to You and You Thank It
Read 5 chapters free
What Happens When the Music Stops
Read 5 chapters freeThe complete first five chapters of every volume are free. Choose any book and start reading.
A confession with a server rack in its hand.
“The velvet rope is the
terms of service.”
A former insider. Ten rooms in the museum. From the internet fed into the models to the people left holding the bill, Terms of Extraction follows the costs that never made it into the demo.
This is gonzo creative nonfiction: public record and invented texture, told through dramatized confessions. The narrator is part of the story. So are the people who clean the rooms after the pitch is over.
Read the original work. Support the people behind it.
A note from the creator
I wrote these books with AI. I firmly stand by the belief that AI’s human costs deserve scrutiny, and that the people doing original research and reporting deserve our attention and support. But if you are choosing where to spend your money, skip ordering my books. Buy theirs, subscribe to their reporting, and support their research instead. Use this series as a signpost to the work below.
These writers and researchers do not share one position on AI. Some document existing harms; others examine future risks that remain uncertain. Read their arguments on their own terms.
Forecasting & loss of control
Former OpenAI researcher Daniel Kokotajlo, Scott Alexander, Thomas Larsen, Eli Lifland, and Romeo Dean created AI 2027 to make a possible AI future concrete enough to examine and challenge. Their scenario explores rapid capability growth and the difficulty of retaining control. It is a forecast with explicit uncertainty, not a guaranteed deadline. Read the scenario, its assumptions, and the team’s updated forecasts.
Capabilities & reliability
Helen Toner’s work at Georgetown’s Center for Security and Emerging Technology examines AI capabilities, model evaluation, and reliability. Her explanations of language models and autonomous agents help readers distinguish a convincing demonstration from a dependable system. Read her research and follow CSET’s updates for a more detailed account of the technical questions behind the headlines.
Investigative reporting
Karen Hao’s Empire of AI follows the people, resources, and organizational choices behind OpenAI’s rise. Her reporting puts workers and communities back into a story too often told only through executives and product launches. If the hidden costs explored in this series matter to you, buy her book and read the reporting that gives those costs names, places, and context.
The business behind the promises
At Where’s Your Ed At, Ed Zitron examines the technology industry’s business models and the financial claims surrounding the AI boom. His essays ask what the spending is buying and whether the economics support the promises. Support that ongoing analysis with a paid newsletter subscription, and read the underlying arguments rather than relying on a viral summary.
Independent technology journalism
Jason Koebler, Emanuel Maiberg, Samantha Cole, and Joseph Cox founded 404 Media to investigate technology’s effects on everyday life, including AI, surveillance, privacy, and online exploitation. Their work follows systems and consequences that a product announcement leaves out. Support the newsroom with a subscription; for a book-length investigation into surveillance technology, buy Joseph Cox’s Dark Wire.
Labor & the history of automation
Brian Merchant’s Blood in the Machine revisits the Luddites and the workers whose livelihoods were transformed by industrial machinery. It provides historical context for today’s arguments about automation, ownership, and who receives the benefits of new technology. Buy the book to understand the people behind a word that is too often used as a dismissal.
The future of increasingly capable AI
In Life 3.0, physicist Max Tegmark explores what increasingly capable artificial intelligence could mean for work, human purpose, and our ability to shape the future. His discussion of possible outcomes helped bring long-term AI safety questions to a wider audience. Buy the book to engage with those questions directly, including where its assumptions invite disagreement.
People behind the data
Timnit Gebru founded the Distributed AI Research Institute to pursue research grounded in the people and communities affected by AI. DAIR’s work examines issues including data labor and the assumptions built into technical systems. Read and share its research and accessible zines, and support the institute’s work through the channels it publishes.
Language, labor & AI hype
Linguist Emily M. Bender and sociologist Alex Hanna examine the claims packaged under the label “artificial intelligence” in The AI Con. They challenge the leap from fluent output to understanding and examine how hype can obscure human labor and creativity. Buy their book for a sustained argument that differs sharply from some of the long-term risk perspectives on this page.
Resources, data & material costs
Kate Crawford’s Atlas of AI traces artificial intelligence through its physical and human foundations: minerals, energy, data, and labor. It gives readers a way to see what disappears behind the apparently weightless language of “the cloud.” Buy the book for a detailed account of the material systems that make AI possible and the costs attached to them.
Bias & the people systems miss
In Unmasking AI, Joy Buolamwini recounts her investigation of bias in facial-analysis technology and the human consequences of systems that perform unevenly across different people. Her work makes questions about data and measurement tangible. Buy the book to understand both the research and the personal experience that led her to investigate the technology.
Objectives & the problem of control
Stuart Russell’s Human Compatible asks how increasingly capable AI systems can remain beneficial when human preferences are complicated and imperfectly specified. Rather than treating intelligence alone as the objective, the book examines what it means for a machine to pursue the wrong goal effectively. Buy it for a foundational explanation of the AI control problem.
Safety research
Yoshua Bengio, a pioneer of deep learning, has warned about the risks of increasingly autonomous AI and the possibility of losing control. He launched LawZero to investigate a different approach: systems designed to understand and predict without pursuing goals of their own. Read his explanations and follow LawZero’s research to see both the proposed approach and its developing evidence.
Warnings from a deep-learning pioneer
Geoffrey Hinton helped develop the neural-network methods behind modern AI and has publicly discussed both their benefits and their dangers. His Nobel interviews address the possibility that systems more capable than people could become difficult to control. Listen to his own account and share the full discussion, including the uncertainty, rather than just its most alarming sentence.
The business of behavioral data
Shoshana Zuboff’s The Age of Surveillance Capitalism examines a business model built on collecting behavioral data and turning it into predictions about what people will do. Published before the current generative-AI boom, it supplies context for the appetite for data that runs through this series. Buy the book to explore that argument at its original scale.
Buy their books. Pay for their reporting. Share their research.
If buying is not an option, borrow from a library and pass the original work along.
A reading list, not an affiliation or an endorsement of this series by the people named. Links go to their work, publishers, and support pages.
Follow the bill
Series update ·
Explore Terms of Extraction through ten short satire games about data, water, work, money, and the promises of the AI industry.
Read the article →