Episode Summary
Executive Summary: The episode centers on a long, contentious interview with New York Times reporter Cade Metz about his viral Slate Star Codex story, journalism ethics, anonymity, and the changing role of tech media, then shifts to an extended discussion of Metz’s new book on AI. The conversation explores AI’s origins, the rise of rationalist culture, algorithmic amplification, misinformation, China’s AI ambitions, and the potential risks and benefits of advanced AI.
Main Topics: Slate Star Codex controversy and naming the anonymous blogger: The hosts debate why the NYT story identified the pseudonymous blogger Scott Alexander, whether that was a journalistic necessity, and how Metz assessed safety concerns and accusations of doxxing. Journalism objectivity vs advocacy and the NYT’s role: A long back-and-forth contrasts old-school neutral reporting with modern opinionated or advocacy journalism, including the impact of Trump-era polarization and the blurred line between news and opinion. Tech media, Silicon Valley, and changing coverage norms: They discuss how tech once received mostly favorable coverage, how adversarial outlets like The Register shaped reporting, and how the tech industry increasingly sees mainstream journalism as hostile. Rationalists, the Overton window, and extremist adjacency: Metz explains that Slate Star Codex and related communities serve as a window into a mindset that wants to discuss any idea, even extreme ones, which critics argue can normalize racism, eugenics, and other harmful views. Algorithms, engagement, and misinformation: The conversation broadens into how social and video algorithms reward emotionally intense content, create rabbit holes, and can push users toward more extreme material or misinformation. AI history, talent, data, and compute as the new moat: Metz summarizes the book’s core thesis: modern AI emerged from a small set of academic breakthroughs, then became an industrial race defined by data, talent, and massive compute resources. AI risks: bias, surveillance, weapons, and national security: The discussion covers AI bias in data, surveillance uses in China, autonomous weapons, quantum computing concerns, and the possibility of AI accelerating both useful science and harmful state power.
Key Arguments: Metz argues his job is to report facts and context, not to advocate, and that the Times newsroom is distinct from its opinion section. The host argues that journalism has become more partisan, more click-driven, and more willing to blur reporting with activism. Metz says the tech industry enjoyed years of favorable coverage before scrutiny increased, and that this shift was overdue. The Slate Star Codex story was framed as a report on a specific influential subculture, not an attack on all of Silicon Valley. Metz maintains that Scott Alexander was already publicly identifiable in the community and had published under his own name elsewhere, weakening the claim of doxxing. The rationalist ecosystem is presented as a place that encourages airing controversial ideas; critics fear that this normalizes extremist views or creates a pipeline to them. Algorithms, especially recommendation systems, are argued to amplify the most emotional or engaging content rather than the truest content. AI is presented as powerful because it scales pattern recognition across speech, images, search, and biology, but that same power can intensify bias, surveillance, and misinformation. The book’s central historical claim is that a handful of researchers, especially Geoff Hinton and Demis Hassabis/DeepMind, turned neural networks from fringe ideas into the foundation of modern AI. China is highlighted as a major AI competitor with distinct political and surveillance incentives, making AI a strategic national-security issue.
Data Points: Age of podcast host and guest generation: Both described as Gen X / around 50 - Used to frame shared old-school journalism training and perspective. Length of rationalist blog posts: 8,000 words long - The host describes Slate Star Codex posts as unusually long, dense reading. Times since he had heard of the blog before controversy: About 4 times - The host says he had only heard of the blog a few times before the scandal. LinkedIn members worldwide: Over 722 million - Mentioned during the LinkedIn Jobs ad read. Vanta offer: $1,000 off - A promotional discount for SOC 2 compliance services. DeepMind Go breakthrough year: 2016 - Metz recalls the Seoul match where AlphaGo defeated Lee Sedol. Lee Sedol result: A machine beat one of the best Go players in the world - Referenced as an AI inflection point. DeepMind protein-folding breakthrough: End of 2020 - Metz says AlphaFold solved the protein folding problem. OpenAI funding mentioned: $1 billion - Discussed in relation to compute needs for modern AI systems. AI talent compensation: Low millions of dollars annually - Metz says top AI talent can earn one, two, or three million dollars per year. Go board size: 19 by 19 - Explained to show why Go is much more complex than chess. Go positions: 361 intersections - Used to illustrate the huge state space of the game. Podcast episode count references: Episodes 573, 673, 808 - The host notes Metz’s prior appearances on the show.
Pivotal Quotes: "My aim when it comes to writing a story about this or writing a book, I am not a player." — Cade Metz: Explaining his reporting philosophy and claim of neutrality. "This story is about the third thing when you thought there were only two." — Cade Metz: Describing the rationalist / Silicon Valley mindset as neither simply left nor right. "We all have to think about this stuff." — Cade Metz: On unintended consequences of technology, algorithms, and AI.
Implications: The episode underscores how tech reporting, platform algorithms, and AI are now intertwined with politics, safety, and power. Listeners are left with a warning: the same systems that accelerate innovation can also amplify extremism, bias, surveillance, and misinformation.
About This Week in Startups
Jason Calacanis covers startups, tech, markets, media, and all the hottest topics in business and technology. He also interviews the world’s greatest founders, operators, investors, and innovators.