Exclusive Interview with a16z Founder Chris Dixon: The Intersection of Artificial Intelligence and Cryptotechnology
TechFlow Selected TechFlow Selected
Exclusive Interview with a16z Founder Chris Dixon: The Intersection of Artificial Intelligence and Cryptotechnology
Supporting new architectures (such as blockchain) and open-source AI is crucial. This is not only a technical issue, but also a regulatory and public perception issue.
Navigating Web3 tides with focused insightsAuthor: a16z
Translation: CaptainZ
Welcome to this special Web3 episode presented by a16z. This episode focuses on the convergence of artificial intelligence (AI) and crypto, featuring Chris Dixon, founder and managing partner of a16z Crypto, and David George, partner at a16z Growth Fund.
They will dive deep into topics such as flaws in the internet's economic model, new opportunities for creators, and the profound impact of shifts among major platforms.
This content stems from a cross-collaboration within a16z’s "Conversations on the AI Revolution" series, coinciding with the paperback release of Chris Dixon’s bestselling book Read Write Own, making it especially significant. For more information, please see the links in the show notes.
Please note that this program does not constitute tax, business, legal, or investment advice. Visit a16z.com/disclosures for additional important information, including our investment disclosures.
Below is the full transcript:
Host: Chris, thank you for joining us. It’s great to have you here. You’re currently focused on crypto—can you share your overall view on how AI and crypto intersect?
Chris Dixon:
Sure, happy to be here. I've always believed that technological waves tend to come in pairs or groups. Over the past 15 years, cloud computing, mobile internet, and social networks reinforced each other. Mobile put computing devices in billions of hands, social networks became the killer app for user engagement, and cloud computing provided the infrastructure. All three were essential. I remember people arguing about which was most important, but they turned out to be deeply interdependent.
Now I believe AI, crypto, and new hardware (like robots, autonomous vehicles, VR headsets, etc.) are forming the three pillars of a new wave of technology that will similarly reinforce one another. Crypto—which is a central theme of my book—offers a fundamentally new architecture for internet services. It’s not just a technology, but a new paradigm for building networks, with properties unachievable through traditional approaches. I think this benefits many domains.
Many people equate crypto simply with Bitcoin or meme coins, but myself and many smart people in the field see it as much deeper than that. The intersection of AI and crypto takes several forms. First, a straightforward connection—one of our key investment themes—is using this new architecture to build AI systems. A core issue in AI today is whether AI will be controlled by a few large companies or by broader communities. This ties into open source. I’ve been surprised to see AI shift over the past decade from being fully open (papers public, code shared) to increasingly closed. Large companies lock down technology citing “security,” but I suspect commercial interests play a bigger role than genuine safety concerns.
Luckily, there are still open models like LLaMA, Flux, and Mistral—but their openness remains questionable. Model weights aren’t fully public, data pipelines aren’t transparent, and true replication is uncertain. And these projects often rely on a single company, making them vulnerable to sudden shutdowns due to strategic changes. That’s why we’re investing in blockchain-based internet service stacks aimed at creating decentralized, open-source infrastructure for AI. Take Jensen, for example—it crowdsources compute capacity in an Airbnb-like model: startups submit AI tasks to the network, individuals with spare computing power contribute, and blockchain manages matching and economic accounting.
Another example is Story Protocol, which reimagines intellectual property registration. You can create an image, video, or music piece and use blockchain to register its copyright and usage terms. These terms are based on existing copyright law and internationally applicable. You might set rules like “allow adaptations and derivatives, but pay me 10% of revenue.” This creates an open marketplace, replacing traditional models requiring individual negotiations. Right now, only big players like OpenAI can make million-dollar deals with Shutterstock, while smaller creators either get stolen from or ignored. Story Protocol levels the playing field.
The core idea here is “composability”—a common theme in blockchain and a concept I explore in my book. It mirrors the success of open-source software: countless small contributions combine into powerful systems. Linux grew from 0% market share in the 90s to 90% today thanks to this force. Story Protocol works the same way—you could imagine someone creating a character, another adding elements, others remixing, eventually forming a superhero universe. As long as revenue flows back per agreed terms, creators stay incentivized. This model embraces new tech while offering creators sustainable economics—the part of AI-crypto convergence I find most exciting.
Host: That new economic model is indeed thought-provoking. David, you mentioned earlier that ChatGPT might break a certain covenant of the internet—can you elaborate?
Chris Dixon:
Yes, there’s a chapter in my book called “The New Covenant” that discusses exactly this. The internet succeeded largely because of clever incentive structures that got five billion people to voluntarily join without any central authority enforcing it. Over the past 20 years, an implicit economic covenant emerged—especially between social networks, search engines, and content creators. Take Google: website owners allow Google to crawl and display snippets, on the condition that traffic flows back. Creators monetize that traffic via ads, subscriptions, or other models. This mutually beneficial relationship underpins the internet’s prosperity.
But occasionally, this covenant breaks. Google’s “One Boxing” feature, for instance, shows answers directly without linking back to original sites—hurting platforms like Stack Overflow, Wikipedia, and Yelp. User experience may improve, but creators lose traffic. Now, AI further challenges this covenant. Chatbots can generate illustrations or recipes directly, eliminating the need to click through to source websites. If all AI systems operate this way, traffic stops flowing back, undermining creators’ livelihoods.
These AI systems are trained on data generated under the old covenant, yet the new model no longer honors those rules. I worry the internet could become a closed system dominated by three to five companies, causing the rest of the world’s billions of websites to wither. That unsettles me—are we reverting to a 1970s broadcast TV model with just a few channels? What good does that do for startups, innovation, or creativity? How do long-tail sites survive? How do new things break through?
I’m not saying crypto is the only solution, but we must at least acknowledge that the current setup breaks the original incentives—and ask whether that’s desirable. If not, how do we design new ones? Story Protocol is one attempt, using blockchain to rebuild incentives for creators.
Host: You said AI, crypto, and new hardware form a triad that reinforces each other. Can you elaborate on how they work together?
Chris Dixon:
Absolutely. Just as mobile, social, and cloud computing enabled each other, so too will AI, crypto, and new hardware. You’re already seeing early signs: AR/VR glasses and autonomous vehicles rely heavily on AI, and companies like Tesla are advancing humanoid robotics. These technologies bring AI into the physical world, unlocking entirely new applications.
On the crypto side, I’m particularly excited about DePIN (decentralized physical infrastructure networks). Helium, for example, is a community-owned, crowdsourced telecom network challenging traditional providers like Verizon and AT&T. Users install Helium nodes (wireless transmitters) at home to extend network coverage. There are now hundreds of thousands of nodes across the U.S., offering service at a fraction of the cost (around $20/month vs. $70). This works because crypto enables incentive design that avoids the multi-billion-dollar infrastructure costs of traditional operators.
The hardest part of building any network is bootstrapping—early on, network effects are weak. Like a dating site: 10 users is useless, 1 million is valuable. Crypto solves this with token incentives—early participants earn rewards, driving network growth. DePIN isn’t limited to telecom; it extends to climate modeling, map data, EV charging, and more. We recently invested in a decentralized energy monitoring network, and others are applying similar models to decentralized science. These early-stage networks are naturally suited to crypto, and AI can complement them in data collection and processing.
Host: The stages of technological development are also crucial. How do you view AI’s evolution?
Chris Dixon:
I like analyzing tech development through a three-stage framework: first, “skeuomorphic”—using new tech to improve existing things; second, “native”—creating things previously impossible; third, “second-order effects”—profound societal changes after widespread adoption.
Take the internet: in the 90s, it was skeuomorphic—moving magazines and catalogs online. Amazon made buying books easier than flipping through print, but it was still essentially old things in new form. In the 2000s, social networks emerged—truly native applications with no offline equivalent and novel business models. AI follows a similar path. The first phase—now underway—involves skeuomorphic uses, like AI customer service replacing call centers. It’s cheaper and more efficient, potentially affecting tens of millions of jobs, but also creating new opportunities. This stage could last 20 years.
The second, “native” phase is what truly excites me. After photography matured, art evolved toward abstraction (e.g., cubism), and cinema emerged as a new art form. Generative AI today may similarly enable new creative expressions—virtual worlds, new game genres, films, or even entirely new interfaces. These innovations require visionary creators and often arrive unexpectedly. Just as cinema opened new frontiers, AI may do the same.
The third stage is “second-order effects.” After social networks rose, Obama’s 2008 campaign marked a turning point, followed by movements like Trumpism and global populism—these are second-order effects, still unfolding. AI’s second-order effects may take 20–30 years to fully emerge, with each phase lasting roughly a decade.
Host: What limits the transition from the first to the second stage?
Chris Dixon:
In the early internet, physical infrastructure—laying cables—was the bottleneck. For AI, technical capability is no longer the main constraint; it’s human creativity and policy. On the supply side, we need creative talent to build native applications. Today’s startup ecosystem is far more mature than 15 years ago—thousands of VCs instead of dozens, better advice, smarter people entering the space, abundant capital and energy.
But demand-side change is harder. Organizational and personal behavior shifts take time. I’d love to use AI to read my book in my voice, but publishers and Audible ban AI outright due to union rules and tradition. Hollywood may need a generation to embrace AI-native films—or it might take emerging-market AI startups to lead. Policy is even more complex. Regulated sectors like copyright, healthcare, and finance—accounting for 70% of the economy—will face intense debate. Is AI training “copying” or “learning”? That may ultimately be decided by Congress, not markets or courts.
Host: What does your ideal future internet look like?
Chris Dixon:
We’re at a crossroads. The original vision of the internet was community ownership, community governance, with value flowing to small businesses, innovators, and entrepreneurs at the edges. But today, wealth and power are concentrated in a few large companies—five tech giants now dominate over half the market cap. The first sentence of my book is “Architecture determines destiny”—control and money flow depend on how services are designed.
I worry we’re approaching an irreversible tipping point where the internet is monopolized by five companies. They’ve saturated user growth and are now “pulling up the ladder,” blocking newcomers. This poses a huge threat to “little tech.” If startups must pay massive “taxes” to compete, they can’t challenge the status quo. We’ve seen this before—Zynga relied on Facebook and ultimately fell victim to platform risk.
That’s why supporting new architectures—like blockchain—and open-source AI is critical. This isn’t just a technical issue—it’s also about regulation and public awareness. We need policies that encourage competition and innovation, not ones that consume tomorrow’s seeds. Through a16z’s efforts, I’m optimistic that the ethos of “little tech” is spreading, and more people are recognizing the importance of new infrastructure and open source.
Host: Thank you, Chris. Great conversation.
Chris Dixon: Thanks for having me—this was fun!
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
a16z
