Interview
Kevin Systrom: Instagram | Lex Fridman Podcast #243
- The emergence of new social networks is expected to occur every three to five years, driven by technical shifts and the potential for unseen groups, such as students or garage teams, to leverage core technologies like reinforcement learning to disrupt incumbents unable to pivot due to structural limitations.
- Future platforms are predicted to shift from connection-based discovery to content-based "search" mechanics, where algorithms match user queries against vast content libraries regardless of social graphs, potentially utilizing virtual world simulations to optimize feed selection and interfaces before real-world deployment.
- Machine learning, specifically reinforcement learning and self-supervised learning, is anticipated to have massive impacts beyond social media, including optimizing climate solutions like energy consumption and logistics, and achieving more widespread application in self-driving cars and complex simulation environments sooner than in social platforms.
- Business leadership and entrepreneurship will increasingly prioritize optimizing for long-term human happiness and aggregate rewards over short-term engagement, requiring new leadership education in supervised and reinforcement learning to thoughtfully apply these technologies while managing trade-offs that prevent unintended side effects.
- Successful ventures are expected to succeed by focusing on specific "jobs to be done" within market cracks, often bootstrapping with single-player modes to build initial value, while avoiding over-engineering and maintaining a "truth-telling" environment through self-funding or aligned incentives to face reality.
- Founders are advised to explicitly opt into the "hard life" of entrepreneurship by choosing games they love, prioritizing the journey over outcomes like wealth, and mitigating the "we have arrived" syndrome by continually seeking the hardest challenges, while acknowledging that humility and a vulnerable "bedside manner" are critical for maintaining public trust.
- Significant risks include the difficulty of defining value and loss functions that account for second-order effects like community building and long-term retention, the structural inability of established players to innovate without losing their core identity, and the psychological hurdles of founder purpose after major exits.
- Market expectations suggest that companies producing consistent "hits" that delight consumers will maintain goodwill, whereas those chasing competitors' features or ignoring negative side effects will face scrutiny, with the ultimate value of a network defined by its ability to solve a single, core human need.