Patrick Collison — Why Silicon Valley's most talented should leave

Career Advice for People in Their 20s:

• The prevailing advice for people in their 20s to move to San Francisco may be suboptimal for those seeking deep technical expertise.
• San Francisco's culture valorizes "iconoclastic" entrepreneurship and dismissing received wisdom (e.g., the Steve Jobs/Bill Gates archetype) but discourages the multi-decade accumulation of domain-specific knowledge.
• Deep technical breakthroughs often require decades of study and experience (e.g., Herb Boyer inventing recombinant insulin later in his career), which the local startup culture may undervalue.
• Entrepreneurs in San Francisco are held in "excessively high regard," creating a status-oriented environment that may not align with all valuable career paths.
• The optimal strategy is to identify the "gradient of maximal learning" for a chosen discipline rather than following a generic geographic or status-based path.
• In fields like biology, true mastery requires direct experience in top labs and mentorship to learn high standards and problem selection, making pure autodidacticism rare.

Scientific Progress and Research Funding:

• There is no direct linear relationship between R&D spending and scientific output; US R&D spending has increased by two orders of magnitude since WWII without a proportional increase in the number of practicing scientists or research output ratios.
• Funding frameworks that assume a constant elasticity between input (dollars) and output (innovation) are flawed; constraints are often structural, cultural, or organizational rather than financial.
• A survey of Fast Grants recipients revealed that 79% stated their research agenda would change significantly if they had flexible, unrestricted funding, suggesting bureaucratic constraints limit impact more than funding levels do.
• The ARC Institute (co-founded by Patrick and John Collison) operates on a model distinct from the NIH:
  • Funding scientists directly rather than projects.
  • Building in-house infrastructure to allow scientists to access capabilities without maintaining them.
  • Creating career paths for senior scientists who wish to remain in research without becoming Principal Investigators (PIs) or managing labs.
• ARC's model aims to support speculative, high-risk research (e.g., bridge editing) that might be rejected by traditional NIH review boards.

Biotechnology and Dual-Use Risks:

• Advances in gene editing (like the recently announced "bridge editing") allow for the creation of structured "telescopes" to systematically perturb genes and understand complex diseases (e.g., Alzheimer's, cancer) rather than just curing specific monogenic conditions.
• Biotech risks are not primarily driven by the lack of scientific knowledge at the frontier, but by the accessibility of known techniques to malicious actors.
• Mitigating biotech risks requires enhancing defensive capabilities (e.g., better disease detection and treatment), which aligns with the goal of curing current human diseases.
• AI and large language models (LLMs) could exacerbate risks by enabling non-experts to synthesize and disperse pathogens, but the core challenge is managing the distribution of existing capabilities.

Artificial Intelligence and Economic Outlook:

• Forecasting AI trajectories requires humility; the primary unknown is the exact shape and limits of current scaling laws.
• The "adaptability premium" is expected to rise, requiring institutions to have competent individuals and new mechanisms (like Fast Grants) to react quickly to unforeseen crises.
• Financial infrastructure for autonomous AI agents is emerging, characterized by high-frequency, usage-based billing and complex liability questions regarding ownership and responsibility.
• Crypto may play a role in AI agent finance, particularly regarding anonymity and the potential tension with KYC/AML regulations.

Stripe's Business Model and Philosophy:

• Stripe views itself as operating in a sector where "moats" are typically overrated; most businesses and products can be done better, and defensibility is often sociocultural rather than purely structural.
• The core "moat" at Stripe is a culture of deep domain understanding, paranoia about missing details, and a commitment to solving problems that others ignore.
• Fast Grants: A $130M+ initiative launched during the pandemic that awarded rapid, unrestricted grants; it demonstrated that removing bureaucratic constraints can accelerate research significantly.
• Stripe Climate: Launched an Advanced Market Commitment (AMC) called "Frontier," raising $1 billion to guarantee purchases of carbon removal.
  • A survey of participating carbon removal companies found 70-74% attributed their founding to the existence of the Frontier AMC.
  • This demonstrates that inducement effects via AMC can be a powerful tool for creating new markets in sectors with market failures (e.g., carbon removal, certain vaccines).
• Payment Infrastructure:
  • Stripe processes approximately 1% of global GDP ($1 trillion annually), a scale that necessitates "multi-decadal abstractions" in architecture.
  • The company deploys production services roughly 1,000 times per day while maintaining 99.995% reliability (approx. 2.5 minutes of downtime per year).
  • Interchange fees (~2-3%) serve as a necessary distribution incentive to cover credit risk, fraud prevention, and customer support; they have enabled the global adoption of consumer credit.
  • Central bank digital currencies and payment schemes (e.g., PIX in Brazil, UPI in India) represent a new era of reinvented national payment rails, often bypassing traditional card network fees for domestic transactions.

Organizational Culture and Governance:

• Writing Culture: Writing is essential for both organizing internal thoughts and creating an intertemporal record for future readers, allowing for rigorous critique and consistency.
• Internal AI Tools: Stripe built an internal platform for LLM integration, enabling collaborative prompt sharing (e.g., SQL optimization) and robust observability across millions of daily invocations.
• Co-Founder Dynamics: The Collisons emphasize the value of long-term partnerships, noting that significant ventures (Stripe, ARC, Fast Grants) were built with family and close friends, citing complementarity and enduring relationships as key success factors.
• Corporate Longevity: The median age of US corporations is low; the speaker notes that Danish companies (e.g., Novo Nordisk, Lego) often achieve longevity by being controlled by nonprofit foundations that legally enshrine mission statements in their constitutions.

Global Economic Growth:

• Stripe's growth is outpacing the general internet economy, driven by the digitization of global commerce and the ability to extend capital to businesses, thereby accelerating their growth.
• Growth drivers include helping businesses enter new geographies, sell in new modalities (e.g., mobile), and optimize existing supply chains.
• Established businesses (incumbents) are credited with a large fraction of global innovation and consumer surplus, challenging the bias that startups are the primary drivers of progress.