Nat Friedman and Daniel Gross in conversation with John and Patrick Collison

The current phase of the AI singularity remains in an early, slow-moving period defined by human-in-the-loop bottlenecks at research labs. Accelerating progress requires removing these human constraints to enable autonomous systems capable of continuous, self-improving workflows. AI agents, such as those capable of reverse-engineering hardware or managing personal health, represent a shift toward a "golden age of tinkering," where software integrates seamlessly with physical devices. Economically, AI adoption mirrors historical disinflationary trends, though it necessitates new organizational structures for managing token-based budgets rather than traditional headcount. As AI capabilities expand, the industry faces a critical transition from purely metric-driven optimization toward a focus on human-centric design, beauty, and the long-term societal impact of these technologies. Future development will likely favor platforms that support agentic transactions, fundamentally altering how enterprises operate and interact with global markets.