At the Pentagon, OpenAI is In and Anthropic Is Out

Fable Ban Reversed + Dr. Dana Suskind on Parenting With A.I. + Prediction Market Drama

Jul 3, 20261h 6mSummary

In this episode, the hosts analyze the recent regulatory crackdown on artificial intelligence, focusing on the U.S. government’s decision to briefly restrict Anthropic’s powerful models due to cybersecurity concerns. The discussion highlights a growing trend toward a de facto, albeit opaque, licensing regime where the government dictates which companies can release frontier models. The hosts criticize the lack of transparent, rule-based processes, arguing that the current approach creates an unpredictable environment for tech developers and businesses alike. The conversation also addresses the narrative regarding international competition, specifically whether Chinese open-source models are closing the gap with American frontier AI. The hosts debate the implications of these regulations, suggesting that while the government is finally acknowledging the significant risks posed by AI, the current strategy lacks coherence. Rounding out the episode, the hosts welcome pediatric surgeon and child development expert Dr. Dana Suskind. Drawing from her background and experience raising eight children, Dr. Suskind explores the intersection of parenting and technology, offering insights on how to nurture human connection, curiosity, and lifelong learning in an increasingly automated world.

‘The Daily’ and ‘The Opinions’: How A.I. Is Changing Loneliness and Taste

Jun 26, 202659 minSummary

In this episode of Hard Fork, the hosts explore the intersection of artificial intelligence and the human experience through two distinct segments. The first part examines the loneliness crisis among older Americans, featuring an in-depth discussion with journalist Eli Saslow. Saslow details his reporting on a pilot program in Washington State where isolated seniors are provided with an AI-powered robot named ElliQ. The conversation highlights how this technology, designed to be proactive and conversational, manages to bridge the silence in the lives of seniors. While the segment acknowledges the genuine companionship and cognitive benefits that can emerge from these interactions, it also grapples with the underlying sadness of substituting human connection with a machine, ultimately questioning if technology acts as a genuine solution or merely a complex crutch for systemic social isolation. The second half of the episode shifts to the concept of taste within the tech industry. The hosts and guests discuss the recent obsession with taste in Silicon Valley, questioning whether AI will ever possess the subjective, ephemeral, and profoundly human ability to exercise aesthetic judgment. They compare the current state of AI culture to the intuitive nature of human taste, debating whether such a skill can ever be truly synthesized or if it remains inherently beyond the reach of large language models.

‘Hard Fork’ Live, Part 3: Differing Visions of an A.I. Future

Jun 19, 202656 minSummary

In this installment of the Hard Fork Live event series, the hosts facilitate a compelling discussion on the trajectory and future of artificial intelligence. The episode centers on an insightful debate between Daniel Kokotajlo, co-author of the influential report AI 2027, and Sayash Kapoor, a researcher at Princeton and co-author of AI as a Normal Technology. The two experts explore their differing worldviews regarding the speed of AI progress. Kokotajlo argues for the possibility of rapid, transformative shifts, including the full automation of AI research and development, while Kapoor emphasizes real-world bottlenecks, such as data efficiency and reliability, which he believes will slow the diffusion of the technology. Despite these divergent timelines for artificial superintelligence, the guests highlight significant common ground, particularly regarding the need for transparency in AI development and concerns over the dangerous precedent set by companies restricting model capabilities for proprietary reasons. The episode also features guest appearances from other prominent voices in the AI scene and a lighthearted demonstration involving a humanoid robot, underscoring the podcast's commitment to making complex technological developments accessible and engaging for a broad audience.

‘Hard Fork’ Live Part 2: Dylan Field on Standing Out in the A.I. Era

Jun 17, 202631 minSummary

In this live episode of Hard Fork, the hosts sit down with Dylan Field, CEO of Figma, to discuss the evolving landscape of design and technology in the age of artificial intelligence. Field reflects on his journey from a curious college student to leading a major design platform, offering his insights on how AI is reshaping creativity. The conversation explores the concept of taste as a crucial differentiator for human designers and the idea that, rather than replacing creators, AI encourages a more iterative, craft-focused approach. Field also shares his views on the current state of Silicon Valley, specifically the trend of founders using AI to build and prototype rapidly, which he suggests allows CEOs to reconnect with the hands-on aspects of creation. Beyond the professional sphere, the discussion touches on the philosophical concept of hyperstition—the idea that stories and beliefs can influence reality—and how intentionally sharing optimistic narratives about AI could help shape a positive future. The episode concludes with a live musical performance by Dan Powell and his robot choir, providing a creative, whimsical capstone to the live event.

‘Hard Fork’ Live, Part 1: Satya Nadella and Cindy Cohn

Jun 12, 20261h 5mSummary

In this special live installment of the podcast, the hosts bring listeners into the room for a series of candid conversations recorded during the show’s annual live event. The episode highlights three distinct segments: an interview with robot-dog creator Phil Mohan, a discussion with privacy activist Cindy Cohn, and a wide-ranging, featured conversation with Microsoft CEO Satya Nadella. The discussion with Nadella focuses on the evolving state of artificial intelligence and its integration into the modern economy. Nadella emphasizes that Microsoft’s strategy is built on creating an ecosystem rather than just a single frontier model, arguing that true economic progress requires widespread adoption and infrastructure that allows every enterprise to operate at the cutting edge. He also touches on the company's hardware vision—specifically agent-first devices—and shares his thoughts on the necessity of balancing technological innovation with sustainable business models, particularly regarding the future of the Xbox division. Throughout the conversation, Nadella remains optimistic about the potential for human-AI collaboration, stressing the importance of maintaining a political economy where technological, market, and democratic forces act as essential checks on one another to ensure long-term societal growth.

Hot I.P.O Summer + What Is A.I. Doing to Math? + HatGPT

Jun 5, 20261h 4mSummary

In this episode of Hard Fork, hosts Kevin Roose and Casey Newton explore the massive shifts occurring within the artificial intelligence industry as three major companies—SpaceX, Anthropic, and OpenAI—prepare for highly anticipated initial public offerings. The hosts discuss the unprecedented scale of these potential IPOs, noting how they could reshape the San Francisco tech economy and further concentrate wealth. They examine the peculiar structure of these firms, particularly how Anthropic’s ties to effective altruism and its status as a public benefit corporation may be tested by the pressures of public markets and shareholder demands for growth. Beyond the financial landscape, the episode pivots to the intersection of AI and mathematics. The hosts are joined by journalist Kevin Hartnett to discuss the dual nature of recent developments in the field: AI is achieving historic breakthroughs in solving complex mathematical conjectures, yet simultaneously triggering a backlash among academics concerned about the erosion of human-led research. The conversation highlights the tension between embracing automated efficiency and preserving the foundational human elements of rigorous mathematical discovery, setting the stage for a broader discussion on the future of intellectual labor in an AI-driven world.

Interesting Times: Why Are We Still Driving?

May 29, 202658 minSummary

In this episode of Hard Fork, the hosts explore the rapidly evolving landscape of autonomous vehicles. They feature a conversation with Andrew Miller, a transportation policy expert and co-author of The End of Driving, who outlines the potential shift toward a world where self-driving cars become a standard fixture in North American cities. The discussion delves into the macro benefits of automation, specifically the promise of increased road safety by eliminating human error, and the micro benefits of reclaiming the millions of hours currently spent behind the wheel. Miller explains the varying technological approaches—comparing Waymo’s sensor-heavy strategy to Tesla’s camera-based model—and touches on the economic viability of robotaxi fleets. A significant portion of the episode examines the complex hurdles of liability and public trust. The participants discuss why manufacturers are hesitant to assume full legal responsibility for their systems and how "inhuman" mistakes by autonomous vehicles can uniquely disturb the public, even when those systems are statistically safer than human drivers. Ultimately, the episode provides a thoughtful look at the tension between technological efficiency and the societal changes required to embrace a driverless future.

Our Field Trip to Google I/O + A Sit-Down With Sundar Pichai + System Update

May 22, 202655 minSummary

In this episode of Hard Fork, the hosts provide an on-the-ground report from the annual Google I/O developer conference. The discussion explores the company's latest flurry of AI announcements, ranging from agentic search features and advanced video models to the unveiling of the Gemini 3.5 Flash model. The hosts offer their critical perspective on these tools, questioning whether Google’s "spray and pray" strategy—aiming for speed, efficiency, and massive distribution—can effectively compete with the more specialized, cutting-edge models currently dominating the industry. The episode features an exclusive, sit-down conversation with Google CEO Sundar Pichai. During the interview, Pichai addresses the company’s current standing in the competitive AI landscape, candidly discussing areas where Google is still playing catch-up, particularly in agentic coding. He also offers his broader philosophical outlook on the societal impact of AI, arguing that despite public anxiety and the inevitability of job market disruption, the technology will ultimately act as a powerful tool for human productivity and professional enrichment. The hosts wrap up the episode by reconciling the conference's ambitious technical promises with the practical reality of how these tools will actually function in users' daily lives.

A.I. Safety Is So Back + Mythos Mayhem with Nikesh Arora + Hot Mess Express

May 15, 20261h 7mSummary

In this episode, the hosts analyze a significant shift in the federal government’s posture regarding artificial intelligence. After years of the current administration dismissing AI safety as unnecessary fear-mongering, recent developments suggest a major turnaround. The discussion focuses on how new, highly capable models like Anthropic’s "Mythos"—which has demonstrated an alarming ability to identify novel cybersecurity vulnerabilities—have forced policymakers to confront the tangible risks of unbridled AI development. The hosts explore the resulting internal incoherence within the administration, where officials are simultaneously courting tech leaders to foster innovation while scrambling to implement rigorous safety and vetting protocols. The episode also highlights the role of Palo Alto Networks CEO Nikesh Arora, who joins the show to provide an industry perspective on the cybersecurity landscape. Together, they discuss the growing frequency of AI-assisted cyberattacks and the challenge of balancing technological progress with the need for national security. Ultimately, the hosts reflect on the irony of the current situation, noting that while the shift toward safety regulation is a positive step, it remains a messy, politically fraught process driven by the rapid, real-world impact of frontier AI models.

Can the U.S. Rein in Prediction Markets? + Joanna Stern on Her Year of A.I. Experiments + Our Producer Goes to Attention School

May 8, 20261h 12mSummary

In this episode of Hard Fork, hosts Kevin Roos and Casey Newton dive into the chaotic rise of prediction markets, examining how these platforms have transitioned from niche academic interests to a pervasive cultural phenomenon. The hosts discuss the frequent insider trading scandals plaguing sites like Polymarket and CalShi, highlighting cases where individuals allegedly leveraged non-public information to profit from military movements or event outcomes. They argue that the current landscape, often described as a regulatory Wild West, necessitates federal intervention to protect market integrity and prevent the exploitation of users. The conversation then shifts to journalist Joanna Stern, who joins the show to discuss her new book, I Am Not a Robot. Stern details her year-long experiment in which she integrated AI into nearly every facet of her personal and professional life, from meal planning and writing to utilizing AI companions. She shares her insights on which technologies are currently useful and which remain purely hyperbolic. Finally, the episode features producer Rachel Cohn, who provides an update on her recent experience at attention school, exploring the challenge of maintaining focus in an increasingly distracted world.