Key Takeaways

Crypto infrastructure has reached a tipping point with payments now costing under 1 penny and completing in under 1 second, while stablecoins process transaction volumes exceeding Visa's monthly throughput, creating a unified global payment network that eliminates traditional financial intermediaries.

AI is disrupting industries in unexpected ways, impacting creative and knowledge workers first rather than manual labor as predicted, with capabilities doubling approximately every seven months and potentially fundamentally breaking traditional internet business models built on search traffic and advertising.

Technology markets follow "winner-take-all" dynamics where first place captures majority market share while second and third places get minimal returns, making it critical for investors to identify and back the best company in each emerging category rather than trying to predict which categories will succeed.

Both crypto and AI represent coordination solutions to different problems—crypto solves collective action and payment coordination challenges while AI creates intelligent systems—and both technologies share characteristics of continuous performance improvements and dramatic cost reductions.

Successful founders possess "earned secrets" from deep industry experience and demonstrate cross-disciplinary knowledge spanning technical, product, and business domains, with the best startup ideas often appearing like bad ideas initially since obvious opportunities are already being pursued by large companies.

Deep Dive

Crypto Infrastructure and Market Evolution

The conversation begins with an assessment of how dramatically crypto and blockchain infrastructure has improved. Core blockchain infrastructure now enables money transfers for under 1 penny and in under 1 second, representing a fundamental shift in payment capabilities. Stablecoins have experienced explosive growth, with usage reaching trillions of dollars—higher than Visa's monthly transaction volume.

The speakers frame stablecoins as a network innovation comparable to WhatsApp's disruption of SMS, creating a single global unified payment network that removes multiple financial intermediaries. This infrastructure enables low-cost, fast international payments and has potential to expand into other financial services like loans, stocks, and treasury bills.

Regulatory challenges significantly hampered progress, with the previous U.S. administration creating substantial obstacles that effectively "lost four years" of potential development and discouraged entrepreneurs from entering the crypto space. However, there's anticipation that upcoming legislation could unlock broader participation, with expectations that more conservative financial institutions will join once regulatory clarity improves.

Practical Crypto Applications and Stripe's Evolution

The discussion turns to real-world implementations, highlighting Stripe's mixed but increasingly enthusiastic relationship with cryptocurrency. Key use cases include cross-border payments and treasury management, with practical examples like:

SpaceX/Starlink moving money between countries
Scale AI paying international contractors

Beyond cost savings, the benefits include fully programmable infrastructure with potential to solve issues like invoice fraud and create reputation and whitelist systems.

AI Development and Unexpected Trajectories

The conversation shifts to generative AI, noting that AI has been in development for 80 years with many unfulfilled promises until recently. While technological progress was somewhat predictable through metrics like GPU improvements, several developments proved surprising:

Generative AI emerged as the leading use case
AI is impacting creative and knowledge worker jobs first, contrary to previous predictions about manual labor
Unexpected sectors being disrupted, including product managers and email-related jobs

The speakers express interest in "second-order effects" of technology, comparing technological evolution to past innovations like automobiles and viewing crypto as a second-order effect of social media.

Social Media's Role in Technology Adoption

Social media enables community formation and evangelization of new technologies (like Bitcoin) that traditional media might initially dismiss. The speakers suggest we're in the early stages of understanding social media's second-order effects on technology adoption.

AI's Creative Evolution and New Media Forms

Discussing generative AI's trajectory, the speakers note that first-order effects include creating illustrations, videos, and potentially feature-length films. Drawing an analogy to photography, they anticipate AI will:

Initially mimic existing media forms (skeuomorphic approach)
Eventually create entirely new media forms that couldn't previously exist

This raises emerging questions about new media forms, business model adaptations, content value in an abundant environment, and copyright implications.

Technology Coordination and Rapid Improvement

The speakers position crypto as solving coordination and collective action problems while AI focuses on creating intelligent systems. Both technologies share characteristics of continuous performance improvements, dramatic cost reductions, and increasing capabilities. Notably, AI models are estimated to be doubling their capabilities approximately every seven months.

Google's Strategic Dilemma and Search Disruption

The conversation addresses potential disruption of established tech giants, particularly Google's innovator's dilemma with its highly profitable but potentially outdated search business model. AI-native activities are showing early signs of challenging Google's search dominance, currently impacting low-monetization areas like knowledge retrieval.

Ironically, Google invented much of the underlying AI technology but faces the difficult strategic decision of potentially cannibalizing its ~$100 billion search revenue. Emerging AI "agents" might enable conducting internet business without traditional search, potentially breaking down the original internet "covenant" between search platforms and content sites.

Internet Business Model Disruption

The discussion centers on how AI (like ChatGPT) might fundamentally disrupt internet business models, particularly websites relying on traffic and click-through rates. Specific concerns include:

AI potentially "boxing" internet content by providing direct answers
Elimination of traditional web traffic monetization strategies
Impact on knowledge-based websites (e.g., Chegg, Stack Overflow)

Key questions emerge about website survival when AI removes the need to visit pages and what new business models might replace current traffic-based revenue. This raises concerns about reducing incentives for creating original content and the need for new compensation mechanisms.

The speakers maintain a hopeful outlook for continued human creativity and meaningful work, anticipating new business models that compensate content creators, similar to how Google and Facebook innovated advertising.

AI Monetization and Competitive Landscape

Predictions suggest AI monetization will likely be more freemium than simple subscription-based. An emerging industry around influencing AI model recommendations (similar to SEO) is developing.

Regarding competitive differentiation, network effects and enterprise selling are considered the strongest competitive moats, though the current AI landscape lacks clear network effects. Despite consumer AI achieving remarkable growth (1 billion monthly active users) with mostly organic adoption, there's uncertainty about the durability of current customer relationships.

Value Concentration and Market Structure

An emerging consensus suggests value will concentrate at two ends of the AI stack: chips/infrastructure and end-user applications. Ecosystem players are working to commoditize the middle layer (foundation models and API serving), with different stakeholders having incentives to commoditize the model layer.

Investment Philosophy and Market Dynamics

The speakers discuss the "Glengarry Glen Ross" market structure, where technology markets often exhibit winner-take-all dynamics. First place captures majority market share, second place gets minimal share, and third place is essentially eliminated. This pattern is considered empirically supported across various tech categories.

Their investment approach focuses on being the "best company in every credible category" rather than predicting which categories will succeed. They prioritize picking the best company within an emerging category through rigorous due diligence, noting that network effects and brand effects are often underestimated in Silicon Valley.

Founder Selection and Startup Dynamics

The conversation concludes with insights on founder selection, emphasizing that good startup ideas often look like bad ideas initially, while obvious ideas are likely already being pursued by large companies. Ideal founders should have "earned secrets" from deep industry or technical experience.

Key founder characteristics include:

Cross-disciplinary knowledge across technical, product, and business domains
Ability to navigate the dynamic "idea maze" of startup development
Deep understanding of their specific market/space
Demonstrated ability to make trade-offs between different aspects of a startup

The investment decision process involves extensive background research beyond single meetings, requiring the investor's own deep knowledge of the industry sector to properly evaluate founders' integrated skill sets and strategic thinking.

Building the Next Internet- Where Crypto Meets AI