Kevin Roose is trying to place the current artificial intelligence moment in a longer, clearer frame. His new book, The AGI Chronicles, and his new media company with Casey Newton both start from the same premise: generative AI is too important to treat as either empty hype or inevitable salvation.
The result is a public argument for AI realism. Roose is not asking audiences to accept every claim made by the industry. He is also pushing back against the idea that the technology is fleeting, trivial, or destined to vanish from the economy without consequence.
A recent history of the AI race
For many people, the current AI Era began around the end of 2022, when OpenAI released the first consumer version of ChatGPT. The product arrived with major attention and, within a few short months, passed over 100 million downloads.
But the story behind today’s artificial intelligence systems is much older. The source article notes that AI has a history stretching back decades, while the work behind current generative AI products has been more than a decade in the making.
The AGI Chronicles focuses on that more recent period. Roose tells the story through three generative AI leaders: OpenAI, Anthropic, and Google. The book is presented as a historical account of the people, institutions, rivalries, and internal drama that helped bring Silicon Valley to its present AI moment.
Roose, described as a longtime tech journalist who most recently served as a columnist and podcast host for The New York Times, conducted more than 150 interviews with AI insiders for the book. His stated hope is that the work can stand as a historical document: detailed enough to explain how the current era took shape, and grounded enough to outlast the day-to-day noise around the technology.
Why he sees a gap in AI media
Roose and Casey Newton recently announced Machine Gods Media, along with a podcast also called Machine Gods. Newton previously cohosted the New York Times’ Hard Fork podcast with Roose.
The new show is being built around a specific critique of the AI coverage landscape. Roose says there is no shortage of podcasts, YouTube shows, or mainstream media coverage about artificial intelligence. In his view, the problem is not volume. It is the shape of the conversation.
On one side, he sees a popular form of criticism that treats the AI boom as fake, hype, or a giant financial bubble. In that view, the tools are not being used, will not affect the economy, and companies such as OpenAI and Anthropic are headed for collapse.
On the other side, he sees coverage that is almost purely promotional. That version of the conversation highlights productivity gains and enterprise use cases while paying little attention to risk.
Roose says the space between those poles is where he and Newton want to operate. He describes this middle position through Newton’s phrase, AI realism: taking AI seriously, recognizing that the tools can be powerful and impressive, acknowledging that they can also be dangerous, and helping audiences understand the subject without turning coverage into boosterism.
The NPR partnership is about reach
Machine Gods will be published in partnership with NPR. Roose says he and Newton value NPR’s broad independent reach and mandate, and that the arrangement gives them ownership of the show and creative control.
That distinction matters to the project. Roose describes the deal as a distribution partnership rather than a full acquisition, leaving Machine Gods Media with operating distance while giving the podcast a path to audiences beyond the usual tech-media circle.
He also emphasizes radio. The source article notes his view that people underestimate how large radio, and especially public radio, still is. Roose imagines a listener encountering the podcast through a local NPR member station, including someone who works in policy, local government, or simply approaches AI from a different point of view.
That public reach is central to the premise. Roose says he does not want the show to serve only an opt-in, self-selected tech audience. If AI is a major public story, the coverage has to travel beyond people who already follow every platform shift, model launch, and Silicon Valley dispute.
Talent, institutions, and independence
The interview also frames Machine Gods Media as part of a broader shift in media. Roose spent nine wonderful years at The New York Times during his most recent stint and says he has spent the vast majority of his career inside large media institutions.
He does not describe the move as ideological. Instead, he presents it as an exciting opportunity and points to a larger audience trend: at least some people want to connect with individuals more than institutions.
That shift creates pressure for traditional media companies. Roose argues that organizations trying to retain and attract very talented people will need to be more flexible about arrangements that let individual journalists build independent brands, shows, and businesses.
The financial side of that independence is also part of the conversation. A Bloomberg report said Roose and Newton were fielding offers of up to $5 million for the show. Roose says the NPR offer is not $5 million, but that NPR provided value beyond money, including radio distribution, a commitment to journalistic excellence, and help growing the show.
What AI realism means in practice
The through line between The AGI Chronicles and Machine Gods is not simply that both are about AI. It is that both treat artificial intelligence as a subject that requires context, skepticism, and seriousness at the same time.
Roose’s book looks backward, documenting how OpenAI, Anthropic, and Google became central to the generative AI story. The show looks outward, trying to make that story legible to a broader audience as it continues to unfold.
That combination reflects the core challenge of AI coverage now. The tools are widely discussed, often misunderstood, and surrounded by intense claims from both critics and promoters. Roose’s answer is to build media that can explain power and risk without collapsing into dismissal or salesmanship.