Altman and Musk hold a dinner to discuss AI existential risks and decide to found a new AI organization.

OpenAI launches as a nonprofit with Altman and Musk serving as co-chairs.

OpenAI restructures to include a for-profit limited partnership designed to raise large-scale investment funds.

Altman leaves Y Combinator and becomes the full-time CEO of OpenAI.

Microsoft invests $1 billion in OpenAI after a GPT-2 demonstration impresses Bill Gates.

OpenAI releases the GPT-3 API despite its own Safety team's warnings about the technology's potential dangers.

Dario Amodei, head of OpenAI's Safety team, resigns and leaves to found a competing AI company.

OpenAI rushes ChatGPT to market to beat competitor Anthropic, with no expectation of its explosive popularity.

ChatGPT reaches 100 million users in two months, making OpenAI a household name worldwide.

Sam Altman testifies before Congress about AI's potential while Hollywood workers protest AI harms in the same building.

The OpenAI board secretly resolves to fire Altman and install Mira Murati as interim CEO.

The board publicly announces Altman's firing, triggering an immediate backlash from employees and tech industry leaders.

Microsoft publicly offers to hire Altman and any OpenAI employees who quit in protest.

Altman is reinstated as CEO after the exhausted board agrees to a partial shakeup.

Investors demand OpenAI convert to a for-profit company within two years or return their funding.

Sam Altman and Elon Musk co-found OpenAI in 2015 as a nonprofit meant to develop AI safely for humanity's benefit. To fund the enormous computing costs required for their ambitions, Altman restructures the organization to include a for-profit arm and recruits Microsoft as a billion-dollar backer. OpenAI's GPT models rapidly improve, culminating in ChatGPT's explosive launch, which makes the company a global phenomenon overnight. Altman's public influence grows alongside OpenAI's commercial success, but behind the scenes his pattern of dishonesty and manipulation strains his relationship with the board. In November 2023 the board attempts to fire him, only to be overwhelmed by a wave of employee and industry support that forces his reinstatement within days. As the book closes, investor pressure pushes the once-nonprofit organization toward a full for-profit conversion, completing its transformation from idealistic research lab to commercial empire.

OpenAI's founding, rise to dominance, and the failed CEO coup.

Karen Hao visits OpenAI's offices to write a profile and finds access severely limited and answers deliberately vague.

OpenAI restricts the release of GPT-2 to select professionals rather than making it publicly available.

Karen Hao publishes a critical profile of OpenAI in MIT Technology Review, triggering internal controversy at the company.

Google fires AI ethics researcher Timnit Gebru after she refuses to retract her paper on large language model dangers.

A partner startup uses GPT-3 API to generate child sexual abuse material, exposing dangerous misuse of the technology.

OpenAI outsources traumatic content moderation work to low-paid Kenyan workers through the third-party contractor Sama.

Annie Altman publicly posts on Twitter alleging emotional and sexual abuse by Sam Altman and her family.

OpenAI's image generator DALL-E 2 launches to public enthusiasm despite internal researcher warnings about misuse risks.

Karen Hao interviews Venezuelan data worker Oskarina Fuentes, who describes years of exploitative conditions on AI labeling platforms.

Mophat Okinyi, a Kenyan data worker, develops severe psychological distress after months of reviewing traumatic content daily.

Sama abruptly ends its OpenAI contract, leaving Kenyan data workers suddenly unemployed without warning or compensation.

A New York magazine profile of Sam Altman brings Annie Altman's abuse allegations to wide public attention.

Altman begins falsely telling associates his sister has borderline personality disorder to undermine her credibility.

Scale AI shuts down all its Kenya operations without warning, stripping data workers of their income overnight.

Annie Altman files a civil lawsuit against Sam Altman alleging sexual abuse beginning when she was three years old.

Karen Hao's investigation begins with her own difficult access to OpenAI's secretive operations, leading her to look beyond Silicon Valley for the full story. She documents how OpenAI and other big tech firms exploit data workers in the Global South — particularly in Kenya and Venezuela — paying poverty wages for psychologically devastating work. The firing of Google AI ethics researcher Timnit Gebru reveals how big tech suppresses criticism of its practices, while the abuse of GPT-3 to generate child sexual abuse material exposes the dangers of releasing powerful models without adequate safeguards. Sam Altman's sister Annie becomes a focal point for Hao's broader argument about power and accountability, as her abuse allegations are met with a coordinated campaign to undermine her credibility. Workers like Mophat Okinyi are discarded without compensation when they are no longer useful, and companies like Scale AI vanish overnight from entire countries they once exploited. Hao frames all of these harms as the predictable consequences of an AI development model that treats people as raw material.

The hidden human costs behind OpenAI's technology.

Ilya Sutskever convinces OpenAI's leadership that massively scaling up computing power is the only path to achieving AGI.

Elon Musk grows dissatisfied with OpenAI's progress, withdraws his funding, and moves his AI ambitions to Tesla.

Researcher Emma Strubell publishes a paper revealing the enormous energy costs of training large language models.

Timnit Gebru and Emily Bender write "Stochastic Parrots," a landmark paper identifying four major ethical risks of large language models.

Altman creates the OpenAI Startup Fund but does not disclose to the board that he personally owns it.

OpenAI launches Codex, a code-generating model built with GitHub and Microsoft, creating a significant new commercial revenue stream.

Greg Brockman scrapes YouTube videos to gather GPT-4 training data in violation of the platform's terms of service.

OpenAI holds a company retreat where Brockman boasts that AI helped his wife diagnose a complex medical condition.

Microsoft agrees to invest $10 billion more in OpenAI after GPT-4 passes demanding science reasoning tests.

Activists in Chile discover Google buried its massive planned water usage in a 347-page document and successfully block the project.

Mira Murati secretly meets with board member Helen Toner to report serious concerns about Altman's disruptive leadership patterns.

Ilya Sutskever contacts Helen Toner and says "abuse" is the word that best describes his own observations of Altman.

Altman lies to a board member, falsely claiming that Sutskever wants Helen Toner removed from the board.

Sutskever withdraws his support for the board coup under overwhelming employee pressure and publicly backs Altman's return.

Jan Leike, leader of OpenAI's Superalignment safety team, resigns and joins rival company Anthropic.

The seeds of OpenAI's dominance are planted when Ilya Sutskever persuades leadership that scaling compute is the only path to AGI, a conviction that shapes every major decision that follows. Early warning signs emerge as Elon Musk abandons the project, critical research is suppressed, and revelations about concealed financial arrangements and exploitative data practices accumulate. Key turning points — including the hidden ownership of the Startup Fund and the scraping of YouTube data in violation of platform rules — reveal how OpenAI pursues growth with little regard for transparency or ethics. As internal tensions peak, board member Helen Toner becomes a crucial figure, receiving urgent warnings from both Mira Murati and Ilya Sutskever about Sam Altman's manipulative leadership. Altman's response — lying to pit board members against each other — exposes his methods plainly, but Sutskever's reversal under employee pressure ensures the coup attempt collapses. The departure of safety advocates like Jan Leike signals that OpenAI has decisively prioritized commercial scale over the founding ideals it once claimed to protect.