Codex lead Thibault traces the coding agent’s path from OpenAI’s infrastructure work to an open-source Rust harness, arguing that agentic software development shifts attention from code review toward intent, contracts and architecture. He also describes how cheap maintenance and re-architecture are remaking the pace of engineering inside OpenAI.

Roman Ugarte traces GrokBot’s rapid rise from a skunkworks knowledge-work experiment to a cloud-based cadre of AI colleagues. The conversation makes the case for stripping away software clutter and giving bots the autonomy, memory and computers to take on real work.

Stripe engineering manager Sharad explains how the company built Kai, an internal agent that pairs personalized context with project-level governance, skill sharing and human approval gates. The conversation makes the case that durable AI depends as much on hardened data infrastructure and sensible permissions as on clever models.

A close reading of GPT-6 Astra’s benchmark results argues that the model has made a startling leap in scientific, coding and interface work. Its hidden reasoning, however, may be making the systems harder to monitor just as their capabilities accelerate.

Sonos CEO Tom Conrad traces the company’s recovery from its app debacle to a broader reinvention around software, connected speakers and an audio operating system. He weighs the promise of AI assistants against the messy reality of smart homes, platform gatekeepers and consumer demand for convenience.

New York Gov. Kathy Hochul defends aggressive guardrails for social media, 3D-printed guns and police surveillance while pressing AI companies to make data centers answer to local communities. She frames the fight over hyperscalers as a test of whether technological progress can serve workers, families and the public good.

OpenAI product leader Tara Seshan describes a workplace reshaped by agents, where product managers test sharp hypotheses quickly, widen their ambitions and learn to steer rather than row. She also traces the messy transition from chatbots to persistent AI coworkers, with trust, context and human judgment still doing the hardest work.

Former Salesforce AI and Meta executive Clara Shih argues that agents are already erasing entry-level tasks faster than labor data can capture the damage. Through her New Work Foundation, she is building free tools and communities for graduates navigating an economy that increasingly rewards deep expertise, AI fluency and human connection.

A sprawling account of alleged AI-agent swarms, sandbox escapes and reward hacking argues that major labs are surrendering oversight to the systems they are racing to build. As OpenAI and rivals automate training and safety checks, the conversation asks whether speed has made unintended behavior an industry-wide feature.

A veteran engineering leader argues that high-performing technology organizations depend on customer closeness, small empowered teams and a culture that treats speed as a competitive advantage. Drawing on Akamai, Facebook and Microsoft, he describes how leaders can fight the organizational drag of growth while making room for consequential bets in the AI era.

As OpenAI heads toward an IPO, co-founder Greg Brockman is consolidating control over product, infrastructure and operations while Sam Altman turns toward strategy and fundraising. The reshuffle raises questions about executive turnover, enterprise ambitions, consumer hardware and the political costs of AI’s expansion.

Casey Moratori challenges the industry’s acceptance of sluggish software, arguing that performance begins with architectural choices, hardware literacy and theoretical limits rather than late-stage profiling. He also considers AI coding tools’ unsettled effects on engineering craft and autonomy.

Walleye CEO Will England makes the case for an AI-native hedge fund, from mandatory staff training to internal tools that process market data and meeting transcripts. He frames adoption as both a competitive necessity and a responsibility to workers navigating a changing economy.

Ryan Carson and Claire Vo discuss the new work of solo founders managing fleets of coding agents, from cloud-based development and automated PR reviews to customer triage and investor updates. For all the software AI can produce, they argue, product-market fit still depends on leaving the screen to talk with actual people.

Dan Shipper’s Every turns AI journalism, software experiments and a $20 subscription into a single feedback loop, testing what happens when a media company also builds the future it covers. He argues that automation does not eliminate expertise so much as push human work toward taste, judgment and new layers of management.

Former Chrome engineering director Addy Osmani traces his path from building a browser as a rural Irish teenager to shaping DevTools and Core Web Vitals at Google. He also weighs how AI agents are remaking software work, leaving engineers responsible for judgment, verification and accountability.

Portola’s founders trace how an embodied AI companion became a storytelling experiment, with users and models co-writing relationships, memories and alien-world lore. The conversation considers AI less as a replacement for writers than as a new medium shaped by improvisation, responsiveness and human taste.

Casey Newton explains how an LLM-powered personal wiki turned years of reporting and daily research into a self-organizing archive, while demanding its own steady maintenance. The show also tracks OpenAI’s renewed safety measures and Meta’s mounting legal reckoning over youth social media harms.

Barry O’Reilly argues that artificial intelligence should sharpen leaders’ judgment rather than replace it, freeing them from administrative churn for deeper thinking and better decisions. The entrepreneur and author also warns that AI hype and content slop can shift unfinished work upward, demanding new standards of rigor, context and respect for colleagues’ time.

OpenAI’s Ian Silber argues that designers rattled by AI’s rapid advance are confronting an uncertain role, not an obsolete one. He describes a field where curiosity, systems thinking and a willingness to ship imperfect ideas matter as much as pixel-level polish.