Where Crypto Meets AI with Chris Dixon & David George

This episode explores the profound economic and structural shifts as AI and crypto converge, reshaping internet business models and creating new opportunities for investors and researchers at their intersection.

The Current State of Crypto

• Chris Dixon, a prominent voice in the crypto space, initiated the discussion by acknowledging the "tricky time" crypto has faced over the last four years, particularly on the regulatory and policy front, which he believes has set the industry back.
• Despite these challenges, Dixon highlighted significant improvements in core blockchain infrastructure.
• A key achievement is the ability to send money for under one penny and in under one second, a stark contrast to the $10 transaction fees a couple of years ago. This is now accessible through platforms like the Coinbase wallet.
• Stablecoins, which are cryptocurrencies designed to maintain a stable value by pegging to a reserve asset like the US dollar, have seen dramatic growth in usage, now processing trillions of dollars monthly, surpassing Visa's volume. Dixon emphasized that this growth is not correlated with speculative trading, suggesting real-world utility.

Stablecoins: The "WhatsApp Moment for Money"

• Chris Dixon described stablecoins as creating a "single global unified network without intermediaries that allow you to make payments," much like WhatsApp unified communication by removing SMS charges and network fragmentation.
• He explained that blockchains and stablecoins essentially offer digital services by removing intermediaries, making transactions lightweight and efficient.
• Dixon believes that once the infrastructure achieves the "one penny, one second" standard, its utility becomes immensely popular.
• He expressed hope for forthcoming legislation, which he anticipates will accelerate stablecoin adoption by allowing conservative financial institutions like banks to participate, thereby enhancing network effects. "Think of it as it's a network effect. And you've got 90 probably five 90 or high percent or some percentage of the nodes on the network that right now don't feel like they can participate," Dixon noted.
• This development is seen as a stepping stone for other financial instruments like loans, stocks, and treasury bills to move onto blockchains.
• David George, sharing Dixon's optimism, pointed to Stripe's recent major product conference, where their largest acquisition ever—a stablecoin company—was a central focus, signaling mainstream adoption.

Stripe's Renewed Interest in Crypto

• Chris Dixon elaborated on Stripe's history with crypto, noting their early involvement with the co-creation of Stellar, followed by a period of disillusionment, and now a renewed excitement.
• Stripe's current use cases for stablecoins include cross-border payments and treasury management, exemplified by companies like SpaceX's Starlink using it to move money internationally.
• Beyond lower transaction costs, Dixon highlighted the programmability of stablecoins as a key advantage. This allows for features like automated reputation systems and whitelisting to combat issues like invoice fraud, where incorrect bank wire information in PDFs can lead to significant losses.
• This programmability is a crucial step towards developing more non-speculative, real-world product use cases for crypto, moving beyond trading applications.

Initial Impressions of Generative AI

• Chris Dixon shared his amazement at the recent advancements in Generative AI (GenAI), a type of artificial intelligence that can create new content, such as text, images, audio, and video. He acknowledged AI's long 80-year development cycle, which had previously led to skepticism.
• Dixon noted the predictability of AI's progress in terms of computational power, referencing Ray Kurzweil's earlier predictions based on Moore's Law-like improvements in GPUs. Kurzweil, a futurist and inventor, is known for his predictions about technological singularity.
• However, he found the emergence of generative AI as the leading use case, over analytical AI, to be surprising. "The fact that it would actually be a creative thing is I think surprising," Dixon remarked.
• The impact on creative professions ("laptop class") before manual labor roles like truck drivers was also an unexpected development.

Second-Order Effects of Technology and AI's Impact on Media

• Chris Dixon discussed the concept of second-order effects, where the indirect consequences of a technology are often more profound than its immediate applications.
• He cited crypto as a second-order effect of social media, arguing that Bitcoin's evangelization and community building would have been impossible without platforms for widespread, non-gatekept communication.
• Applying this to GenAI, Dixon pondered its second-order effects, particularly in media. The first-order effect is the easier creation of existing media like illustrations and videos, with feature-length films and games made by individuals on the horizon.
• He drew an analogy to photography, which initially caused angst among representational artists but eventually led to new art forms like abstract art and, crucially, film—a "native" medium that couldn't exist before photography matured.
• Dixon questioned: "Will there be new forms of media that simply couldn't have existed before [GenAI]?" He hypothesizes that such native forms will emerge, moving beyond skuomorphic applications (designs that imitate real-world counterparts, like a digital notepad resembling a physical one).

New Business Models and the Role of Crypto in the AI Era

• Chris Dixon explored the economic implications of abundant, AI-generated content, suggesting that while content value might decrease, human connection and community around art and creators will remain vital.
• He raised questions about how money will flow and how copyright will function in this new landscape.
• Dixon proposed that crypto can address these challenges, stating, "I think those are all questions that in some ways crypto can address."
• He differentiated AI's focus on intelligence from crypto's focus on coordination and solving collective action problems, such as establishing global payment standards or facilitating capital formation on the internet.
• For Crypto AI investors, this highlights an opportunity for crypto protocols to provide the economic and governance rails for AI-generated content and interactions.

AI's Rapid Advancement and Intersection with Crypto

• David George highlighted the rapid cost reduction in AI, with a 99% decrease over the last two years, and models doubling their capabilities every seven months on certain tasks. This mirrors the cost reduction seen in crypto infrastructure.
• Chris Dixon emphasized that these technological waves—AI, crypto, mobile, social, cloud—tend to reinforce and intersect with each other, rather than compete.
• He suggested that AI's rise resets the "chessboard," potentially threatening incumbents like Google in search and creating openings for new architectures and services where crypto can play a role.
• David George mentioned the concept of AI agents—AI systems capable of conducting business on a user's behalf—getting closer to reality for lower-level internet tasks, further signaling a shift in how internet services are accessed and monetized.

Google's Innovator's Dilemma and the "New Covenant"

• The discussion turned to Google's challenge, a classic innovator's dilemma, where a company's existing successful business model (in this case, highly profitable search advertising) hinders its ability to embrace disruptive new technologies (AI-driven direct answers).
• Chris Dixon pointed out the irony that much of the foundational AI technology was invented at Google, likening it to "Xerox PARC all over again," a reference to Xerox's research center that developed many pioneering computer technologies but failed to commercialize them effectively.
• Dixon introduced his concept of a "new covenant" for the internet. The old, implicit covenant was that distribution platforms (search, social) would send traffic to content sites in exchange for snippets of their content.
• AI, particularly chatbots like ChatGPT, breaks this by providing direct answers, effectively "one-boxing the whole internet." One-boxing refers to when a search engine displays the answer directly within its results, reducing the need to click through to the source website.
• This poses an existential threat to the business models of millions of websites dependent on referral traffic. Dixon urged a discussion on new models for these sites, moving beyond tactical AI model discussions or "cosmological scale" fears of AI.

The Future of Content Creation and Business Models

• David George cited Chegg, a homework help site, as an example of a knowledge-based website immediately disrupted by AI, as its services were largely obviated. This raises the question of who will create the initial content for AI to learn from if the original creators' business models are destroyed.
• Chris Dixon expressed a desire for a "symbiotic relationship" where humans continue to create new genres and ideas, which then feed into and are accelerated by AI systems.
• He hoped AI companies would develop "native" business models, similar to how Google and Facebook created new forms of advertising, rather than just relying on subscriptions. This might involve commerce-based models or new forms of affiliate-like compensation for content sources.
• David George noted that while many recent unicorns are enterprise or freemium, the "best businesses in the world are ad-based" due to free user access subsidized by ads. He bets on more sophisticated, premium models emerging in AI beyond simple subscriptions.
• The potential for a new industry around "convincing the AI model to promote your product" was also touched upon, a new form of optimization akin to SEO.

Market Structures: Network Effects in Crypto vs. AI

• Chris Dixon, drawing from his investment experience, emphasized his focus on network businesses, which are prevalent in crypto. Network effects occur when a product or service becomes more valuable as more people use it.
• He stated, "Crypto is all networks... once it works... there's a very clear defensibility story."
• In contrast, David George observed that current consumer AI applications, despite rapid growth to a billion monthly active users (faster than Google, Facebook, or TikTok, and largely organic), currently lack strong network effects. This makes their customer relationships potentially less durable.
• "There is no network effect. And so, how durable is that customer relationship is something that we wrestle with all the time," George admitted.
• For investors, this distinction is critical: crypto ventures often have inherent defensibility through network effects once established, while AI applications may need to find new moats.

Value Accrual in the AI Ecosystem

• David George outlined the emerging consensus on where value will accrue in the AI stack:
• Clearly at the chip layer (e.g., GPUs).
• Clearly at the end-user application layer.
• The model layer (foundation models served via API) is facing commoditization pressure from cloud companies (wanting to maintain customer power) and application companies (wanting lower costs). Open-source models are also contributing to this pressure.
• Chris Dixon noted that AI applications are still very early, but the ones breaking out are doing so significantly and rapidly.
• The B2B selling motion for AI software is still developing but is expected to follow established patterns of enterprise sales.

Winner-Take-All Dynamics

• David George introduced the "Glengarry Glen Ross market structure" analogy: first prize gets a Cadillac, second prize gets steak knives, third prize is fired. This suggests tech markets often see one dominant winner.
• Chris Dixon largely agreed, stating his firm's investment rule: "the best company in every credible category." They are more humble about predicting *which* categories will succeed but rigorous in picking the *best* company within a chosen category.
• He believes this "winner-take-all" dynamic is empirically borne out, even in cases where obvious network effects aren't initially apparent, citing Uber/Lyft and early Google, and suggesting brand effects are often underestimated.
• "It's better to miss the category than to get the steak knives in my view," Dixon asserted, highlighting the pain of being number two or three.
• For investors, this underscores the importance of identifying and backing potential market leaders, as the spoils are disproportionately distributed.

Picking Founders: The "Earned Secret"

• Chris Dixon shared insights into his founder selection process, emphasizing the concept of a founder with an "earned secret"—deep knowledge gained from extensive work in a specific domain, leading to non-obvious insights.
• He values cross-disciplinary knowledge, as running a company requires integrating technical, product, and business acumen. Founders must navigate the "idea maze"—a dynamic process of making trade-offs across these areas.
• Dixon looks for founders from whom he learns significantly during meetings, indicating deep thought and expertise. Pre-diligence through references and reputation within a space is also key.
• This approach is crucial for Crypto AI investors, as identifying founders who can navigate the complex, intersecting challenges of both fields will be paramount.

Conclusion

• The dialogue underscores that AI and crypto are not just intersecting but are poised to fundamentally reshape internet economics and user experiences. Investors and researchers must track how AI disrupts existing content models and how crypto can provide the coordination and economic layers for new, AI-native ecosystems.