Founder Mode

Jeetu Patel describes how Cisco is positioning itself as critical infrastructure for the AI era and trying to operate with the speed of a startup at the scale of a large company. He explains a leadership approach that combines Cisco operators, outside hires, and founders of acquired companies, alongside a top-down push to get employees using AI. He argues that AI will create new engineering work and drive demand for infrastructure, while making security, observability, and trusted delegation essential. His career lessons emphasize ambitious goals, long-term relationships, compounding progress, and choosing work that fits your ambitions.

Paul Graham argues that the fundamentals of starting a startup stay much the same across technologies: founders need ambition, persistence, and the ability to keep shipping. Ambition is often driven day to day by fear of failure and care for the work, rather than by a distant reward like becoming a billionaire. He describes how AI arrived as a human-like but error-prone system, and says its progress has not changed the core predictors of startup success, though compute costs have become a major expense. He also explains how YC’s batch creates peer support and early customers, and why formidable founders matter more than any particular startup idea.

Mark Pincus argues that consumer products are difficult to fund right now, but AI could enable a new wave of widely used services as intelligence becomes cheaper. He explains his “proven, better, new” method: reuse established product mechanics, improve what users already value, and test one novel idea at a time without becoming attached to it. The conversation also covers founder involvement, company culture, distribution, and how to keep building through periods when no product idea feels compelling. His advice is to start exploring the consumer services that could become possible as AI costs fall, even before they are affordable to everyone.

Harshil Mathur traces Razorpay’s beginnings to the difficulty of accepting online payments in India and explains how conversations with customers led the company to focus on startups. He describes how regulation and patient execution created a barrier to entry, while clear communication during a banking crisis helped preserve customer trust. Razorpay’s early UPI adoption shows how a small company can use speed to enter new markets. He argues that AI will make building software easier and compress technical moats, so companies must move quickly, keep founders involved in product direction, and choose problems they can work on for years.

Derek Thompson argues that an abundant future depends on building more housing, clean energy, infrastructure, and new technologies, and on reforming the institutions that slow them down. He describes how layers of bureaucracy and legal challenges can make well-intentioned policies lose sight of their goals, using rural broadband and California high-speed rail as examples. He calls for renewed science policy, including experiments with longer-term and higher-risk research funding, and discusses how public funding can spur private innovation. He encourages founders and engineers to follow their curiosity toward difficult problems that could reduce pain and increase people’s agency.