Fireside Chat, Interview
Where Crypto Meets AI with Chris Dixon & David George
Crypto Infrastructure & Stability
- Transaction costs have dropped from ~$10 to under $0.01 with settlement times under one second.
- Stablecoin monthly transaction volume has surpassed Visa, reaching the trillions of dollars.
- Stablecoin usage growth is uncorrelated with speculative trading volumes, indicating adoption in real-world use cases.
- Major enterprise players, including Stripe (via acquisition) and SpaceX/Starlink, are adopting stablecoins for cross-border treasury management and payments.
- Programmable money enables automated invoice verification, fraud prevention, and whitelist systems, reducing reliance on manual bank wires.
- Regulatory headwinds from the previous administration stalled development for approximately four years, though infrastructure recovery is underway.
- Future legislation is expected to unlock participation from conservative financial institutions (e.g., banks), accelerating network effects.
Generative AI & Market Dynamics
- AI model capabilities are doubling every seven months, while the cost of inference has dropped 99% over the last two years.
- Contrary to early predictions, the primary displacement is occurring in cognitive roles (product managers, writers) and creative tasks rather than robotic labor.
- Knowledge-retrieval search volume for AI-native agents is moving in the opposite direction of traditional Google queries, though high-value commercial search remains resilient.
- Google faces an "innovator's dilemma" with a $100+ billion search business; moving to a chat-first model risks cannibalizing their primary revenue stream.
- The "internet covenant" (content sites providing traffic in exchange for snippets) is collapsing as AI models act as "one-boxing" the entire internet, removing click-through incentives.
- Disruption is most acute for knowledge-based models like Chegg, which are being obviated by direct AI answers, raising questions about future content creation incentives.
- Second-order effects of AI will likely create new native media forms (beyond skeuomorphic applications) and require new business models, potentially moving toward commerce or affiliate-style revenue rather than traditional ads.
Strategic Intersections & Business Models
- Crypto is positioned to solve coordination and collective action problems (money flow, copyright, value distribution) that arise from the AI revolution.
- Value in the AI wave is expected to accrue primarily at the "barbell" extremes: semiconductor chips and end-user applications, with the intermediate model/API layer facing commoditization pressure.
- Crypto adoption is described as a "second-order effect" of social media, similar to how automobiles led to the development of highway systems.
- Google's historical dominance in advertising is compared to Xerox PARC; their technology invented the AI revolution, yet they struggle to pivot due to the profitability of their existing search model.
- The "winner-take-all" market structure (Glen Gary, Glen Ross) remains the dominant framework, where the #1 player captures the vast majority of value while #2 and #3 struggle.
- Brand effects are increasingly recognized as a critical defensive moat alongside network effects, often underestimated by analysts.
Investment Philosophy & Founder Selection
- Investment strategy prioritizes backing the "best company in every credible category" rather than betting on unproven categories.
- Founders are evaluated on possessing an "earned secret" derived from deep, long-term industry experience.
- Ideal founders demonstrate cross-disciplinary mastery (technical, product, business), navigating the "idea maze" where these disciplines intersect.
- Outsourcing core disciplines (e.g., CTO as a cousin) is identified as a critical failure point; founders must integrate these functions internally.
- The firm avoids competing for second-place positions, preferring to miss a category entirely rather than settle for "steak knives" (mediocre returns).