Who Will Own the Internet? a16z’s Chris Dixon on AI and Crypto

Episode Summary

Executive Summary: The discussion argues that AI, crypto, and new hardware are converging into a new internet era, but the key battleground is economic design: who captures value, who controls networks, and whether creators and startups can thrive. The speakers warn that AI may centralize power into a few systems unless new open, incentive-aligned models—like decentralized compute, IP registries, and proof-of-human systems—are built to preserve creativity, competition, and the long tail of the web.

Main Topics: AI, crypto, and hardware as a reinforcing technology wave (Priority: 5/5): Chris Dixon frames AI, crypto, and new devices as a complementary trio similar to mobile/social/cloud, each enabling the others and creating new categories of products and networks. Open-source AI and decentralizing AI infrastructure (Priority: 5/5): The conversation emphasizes concern that AI development is becoming closed and concentrated, and highlights decentralized compute and open services as ways to broaden access and control. New economic models for creative people (Priority: 5/5): A major theme is how creators can be compensated in an AI world through blockchain-based rights management, remix terms, and revenue waterfalls instead of platforms capturing all value. The internet’s broken covenant and traffic shift (Priority: 5/5): The speakers argue that the old bargain—search platforms index content and send traffic back—is breaking as AI systems answer directly, potentially starving millions of sites of distribution. Crypto as a network-bootstrapping mechanism (Priority: 4/5): Crypto is presented as a tool for solving early-stage network problems by incentivizing participation in physical and digital networks, especially where network effects are weak at launch. AI phases: skeuomorphic, native, and second-order effects (Priority: 4/5): The discussion separates AI’s likely trajectory into replacing existing workflows, creating entirely new behaviors and media, and eventually triggering broader societal changes. Regulation, incumbency, and the risk of centralization (Priority: 4/5): The speakers argue that policy, labor resistance, and incumbent business models may slow AI adoption and shape whether the future is dominated by a few large firms or remains open.

Key Arguments: AI should be embraced, not blocked; the real question is how to redesign ownership and revenue models so creators benefit rather than lose out. Crypto is not mainly about meme coins; it is a new architecture for building networks, incentives, and open services. A few AI companies may increasingly answer user queries directly, which could break the longstanding web covenant that rewarded websites with traffic. Story Protocol-like systems could let creators set remix and licensing terms upfront, with automatic revenue sharing for derivatives. Open-source AI is fragile because weights, data pipelines, and frontier capability development are still controlled by a small number of companies. Decentralized compute, proof-of-human systems, and crowdsourced data/evaluation can support AI without relying entirely on centralized intermediaries. AI’s first wave will be skeuomorphic—replacing jobs and workflows—but the most valuable phase will be native applications and new forms of media. The next major policy fights will likely happen in Congress and courts because AI affects too many jobs and industries to be left to market forces alone. Startups need open-source software and access to distribution; otherwise platform risk and incumbent capture will make it hard to compete.

Data Points: Big AI systems: 3 to 5 - Chris suggests the world may consolidate around a handful of dominant AI systems. Helium cellular monthly cost: $20/month vs $70/month - Used to illustrate the cost advantage of a community-owned telecom network. Open source Linux market share: 0% in the 1990s to 90%+ today - Cited as an example of composability and long-term open-source dominance. OpenAI/Shutterstock data deal: $100 million - Referenced as an example of high-end licensing deals between AI firms and content owners. Waymo safety claim: 7 to 10 times safer than a human driver - Mentioned when discussing adoption of AI in regulated industries. Internet companies in the middle: Top 5 internet companies hold more than half of market cap - Used to argue that value has increasingly concentrated in a few platforms. AI-related jobs at risk: Tens of millions - Described as the scale of jobs likely affected by AI automation, especially middle-laptop knowledge work. Startups / venture firms then vs now: About 10 venture firms then, thousands now - Illustrates how much more mature and crowded the startup ecosystem has become.

Pivotal Quotes: "What will be the economic models for creative people in an AI world?" — Host: Opening framing of the episode’s central question about creator compensation and AI. "Don't stop the inevitable, which is the technology progressing. Lean into it and rethink those models." — Host: A core thesis that AI adoption is unavoidable, so the challenge is redesigning incentives and economics. "your architecture is your destiny" — Chris Dixon: Explains his view that technical architecture determines control, governance, and where money flows online.

Implications: AI is likely to centralize attention and revenue unless creators, startups, and policymakers build open, incentive-aligned systems. The winners may be those who combine AI with crypto-native ownership, distribution, and network design.