Fireside Chat, Interview
Kevin Systrom & Mike Krieger: The Founders of Instagram Reveal their New Project "Artifact" | E981
- The product strategy centers on leveraging machine learning to recommend content from news and web information, anticipating that initial recommendations will be accurate for like-minded groups before evolving into a recommendation engine driven by sufficient user data.
- There is an expectation that the platform will prioritize a simple, "vanilla" core identity to avoid overwhelming users, with complex features ("toppings") added later as the user base grows and provides the necessary data for better recommendations.
- The roadmap involves a five-year horizon to establish trust with publishers, spawn independent publishers, and shift societal perception of algorithms from skepticism to excitement, with the long-term goal of having the product solve fundamental questions of how technology serves users.
- Operational plans include working 16 hours daily to balance intense productivity with family time, maintaining readiness to revert to waking up between 2 AM and 4 AM during crises, and using a bootstrap approach to expand the user base and data set without external pressure to launch prematurely.
- Recruitment strategies rely on the founders' network to source talent, with a focus on assessing candidates within five seconds, prioritizing raw talent and the ability to learn quickly over strict role boundaries for the next few years, and moving quickly to remove employees who do not fit the high-variance startup environment.
- The founders anticipate a constant state of "unbreaking" products and facing significant failure, viewing this suffering as a necessary condition for reward and expecting that "muscle memory" for intense work will return instantly during critical periods.
- Life changes, specifically having children, are expected to alter personal dynamics, necessitate a 20-year horizon for decision-making, and drive a focus on prioritization to only do what matters, while also fostering a desire to balance high intensity with eventual social reintegration after five years.
- Partnerships are viewed as dynamic entities requiring conscious work to survive disagreements, life changes, or differing paces of intensity, with a belief that relationships must "morph" to avoid becoming difficult if one partner begins to "coast."
- Market expectations include the "hedonic treadmill" where satisfaction resets at higher value points, a "zero interest" phenomenon that does not last forever, and the belief that San Francisco's economy will continue to rebound after downturns despite a growing remote work component.
- A significant risk involves the "chicken and egg" dynamic of needing large datasets to make good recommendations while needing a liked product to generate that data, with a painful expectation that a large portion of day-one sign-ups will remain uncertain about the product's value.
- The founders aim to navigate the tension between the social graph and machine learning by using algorithms to work for the user rather than the company, expecting that unconnected content will remain important for conversation and relationship building while a total abandonment of the social graph does not occur.
- Future growth may follow a trajectory of starting as a utility and becoming a social behemoth, though there is a possibility of verticalizing community management, and the founders expect the startup process to be volatile, chewing up and spitting out people through high emotions.
- Long-term views on success and happiness suggest that being slightly random on the margin is optimal for machine learning, that a billion-dollar exit does not meaningfully increase happiness compared to enjoying one's work, and that past patterns in history will help the team avoid or correct mistakes faster than first-time founders.