Fireside Chat, Interview
Who Will Own the Internet? a16z’s Chris Dixon on AI and Crypto
- Technology waves involving AI, crypto, robotics, self-driving cars, and VR are expected to reinforce one another, with the next phase of adoption converging generative AI, crypto, and new hardware platforms like AR/VR glasses and humanoid robotics that are currently barely scratching the surface.
- The AI landscape is predicted to follow three stages lasting potentially a decade each: skeuomorphic (doing old things cheaper), native (creating new forms previously impossible), and second-order societal effects, with the transition to the native phase bottlenecked by human creativity and slow organizational change.
- Open-source AI is expected to remain fragile over the long term as many models do not open weights or data pipelines and can be altered by single large companies, while the broader internet faces a risk of breakdown where search engines and platforms stop sending traffic to original content in exchange for indexing.
- Three to five major AI systems are predicted to control the vast majority of information distribution, potentially turning the internet into a broadcast-style model similar to 1970s television, creating a point of no return where five companies kick away the ladder of innovation.
- New economic covenants and incentive systems are required to layer on existing infrastructure to ensure small businesses and creators capture upfront amounts and downstream royalties, preventing platforms from capturing 100% of revenue and ensuring money flows to network edges.
- Crypto is positioned as a new architecture to enable capabilities previously impossible, including crowdsourced compute layers like Project Jensen, intellectual property registration via Story Protocol, and decentralized physical infrastructure (D-PIN) networks like Helium to solve bootstrap problems.
- Helium is expected to offer cellular services at approximately $20/month by utilizing a crowdsourced network, with potential expansion into climate modeling, mapping, and charging infrastructure, while crypto incentives are also expected to replace centralized approaches like Scale AI for model evaluation and data generation.
- World Coin is intended to provide cryptographic proof of human identity in a world where AI replicates humans, replacing CAPTCHA systems that are becoming AI-proof but not human-proof.
- AI integration strategies will likely favor newcomers who treat AI as a core ingredient requiring a rebuild rather than incumbents applying it as frosting, though consumer AI applications face risks of price competition and copycatting by established players.
- Hollywood's adoption of AI is expected to be delayed by a whole new generation due to union resistance, regulatory hurdles, and workflow integration difficulties, while approximately 70% of the economy in regulated industries like medical and finance will be slow to adopt compared to less regulated sectors.
- AI regulation is expected to play out over a very long time through courts, state legislatures such as California, and Congress, resulting in legislation that attempts to balance creator incentives with the existence of AI.
- A "come for the tool, stay for the network" strategy is viewed as necessary but difficult to execute for AI startups to gain traction against incumbents who may exploit platform risk and build on quicksand without new open internet architectures.
- Composability is expected to drive the success of open source through a "Lego effect" where media assets are combined by users while revenue cascades back to original creators, supported by a new blockchain-based economic model for creative people.
- The native phase of AI is expected to unlock new forms of media and art analogous to film replacing photography, pushing the frontier of human creativity rather than merely replacing existing work, provided the economic model shifts to align incentives.
- Without open source AI and open internet architectures to prevent control by a few dominant entities, the "small tech" sector risks losing its ability to innovate and grow, requiring new internet architectures to ensure competitive services can threaten incumbents.