Search for a command to run...

Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
This episode of This Week in Startups features two groundbreaking companies reshaping the future of technology and logistics. Alex interviews Keller Cliffton, co-founder of Zipline, whose drone delivery service just raised $600 million at a $7.6 billion valuation. Zipline now operates more flights per day than United Airlines and is transforming how goods reach consumers with teleportation-like delivery speeds. (09:00) The episode then shifts to Anastasis Germanidis, CTO and co-founder of Runway, who explains their breakthrough in world models - AI systems that can simulate and interact with reality in real-time, opening massive opportunities in robotics, gaming, and autonomy. (32:12)
Co-founder of Zipline, the world's largest commercial autonomous aircraft company. Under his leadership, Zipline has achieved over 135 million commercial autonomous miles with zero safety incidents and now operates more flights daily than United Airlines. The company has raised $600 million at a $7.6 billion valuation and operates the largest commercial autonomous aircraft factory in the United States.
Co-founder and CTO of Runway, a $3 billion AI company pioneering generative video and world models. He started the company at NYU's Tisch School of Art and has been building frontier AI models for seven years. Runway recently announced their first general world model, GWM-1, which can simulate interactive environments for robotics, gaming, and autonomous systems.
Keller explains why Zipline chose to build an end-to-end delivery service rather than selling drones to companies. (13:49) The aircraft represents only 15% of the total solution - the real value lies in integration, ground infrastructure, unmanned traffic management, and customer-facing applications. This approach allows Zipline to control quality and optimize every part of the customer experience, from the "magical portal" installation to final delivery.
The conversation reveals a significant shift toward American manufacturing capabilities in advanced technology. (04:40) While China produces 300 times more basic drones than the US, America leads in sophisticated autonomous aircraft like Zipline's 60-pound vertical takeoff vehicles with 43 custom sub-assemblies. The key is pushing state-of-the-art technology through to manufacturing scale rather than competing on existing low-tech products.
Anastasis describes how world models solve critical challenges in robotics and autonomy by generating data for rare edge cases that would take years to collect naturally. (41:09) Instead of waiting to capture real-world accident data for autonomous vehicles, world models can simulate millions of edge case scenarios, making AI systems more robust and safer before deployment.
Runway's approach to training world models across gaming, avatars, and robotics simultaneously creates powerful cross-domain improvements. (43:57) Skills learned from avatar interactions enhance human behavior simulation in games, while robotics data improves physical manipulation capabilities across all applications. This generalization approach mirrors successful pre-training strategies in language models.
Jason's founder advice emphasizes that teams unable to ship within 90 days likely never will. (60:09) Having founders who write code eliminates communication overhead and acceleration bottlenecks. The Y Combinator question "Who writes code?" screens for execution capability - teams that can build, iterate, and ship rapidly have dramatically higher success rates than those relying on external development.