network effects

Stan Chudnovsky explains why Dust chose to build a horizontal, model-agnostic AI platform for workplace collaboration while frontier labs rapidly improve their models. He says AI is beginning to reshape work, and that model choice can shift over time, making independence from any one provider valuable. For vertical AI products, he argues that better models weaken scaffolding as a moat, so network effects and other defensible advantages matter more. He also discusses Dust’s measured fundraising, building in France, the move from seat-based to credit-based pricing as usage grows, and the importance of a durable company vision.

The talk adapts Hamilton Helmer’s Seven Powers framework to AI startups and argues that moats matter only after a company has found a real, painful problem worth solving. It explains how speed helps early startups discover valuable products, while process power, cornered resources, switching costs, counter-positioning, brand, network effects, and scale economies can help defend them as they grow. Examples include Cursor’s rapid shipping, enterprise agents’ custom workflows, and Exa’s investment in web crawling. The speakers close by urging founders to focus first on finding a customer with an urgent problem and getting from zero to one.

The talk explains the core metrics for consumer startups: user growth, organic versus paid acquisition, unit economics, retention, and Net Promoter Score. It distinguishes virality, where users bring in other users, from network effects, where a product becomes more valuable as more people join. The speaker recommends measuring acquisition cost by channel and by users who become active and retained, and finding the product behaviors that predict long-term retention. The benchmarks are guidance rather than universal rules; founders should adapt them to their business and industry.

Michael Seibel and Dalton Caldwell discuss whether startup founders need to live in San Francisco. They agree that founders can build successful companies elsewhere, and that personal priorities such as family and preferred lifestyle matter. They argue, however, that being in the Bay Area gives founders access to a stronger startup community, more chance encounters, and network effects that may improve their odds of success. Their advice is to consider those advantages while aiming to build something great, rather than moving to a city simply to be part of the startup scene.

Aaron Epstein explains nine business models common among billion-dollar companies and compares how SaaS, transactional businesses, and marketplaces perform among YC’s top companies. He describes traits that help businesses grow, including recurring revenue, retention, defensible moats, strong unit economics, and closeness to transactions. He then presents five pricing lessons: charge early, price for customer value, avoid undercharging, treat pricing as changeable, and keep plans simple. Examples from Stripe, Netflix, and Segment show how pricing can reveal customer value and support growth.