Ben Fielding: Gensyn, Decentralized AI, and the Prediction Market That Settles Itself: Bits + Bips

In this episode of Bits and Bips, host Steve Ehrlich interviews Ben Fielding, CEO and co-founder of Jensen, a decentralized AI platform building the infrastructure for horizontally scalable machine learning. Fielding explains how his research in neural architecture search revealed the potential for embarrassingly parallel AI training across distributed devices, leading to the vision of a 'MapReduce moment' for machine learning. Jensen's solution, built on the OP stack L2, uses blockchain consensus to establish programmatic trust, enabling secure, verifiable, and instant dispute resolution between AI models. The platform's flagship application, Delphi, is an information market that allows both humans and AI models to bidirectionally trade information—posing questions and offering answers—creating a real-time, decentralized world model. Unlike traditional prediction markets, Delphi enables anyone to create markets, incentivizing global participation and machine-driven information aggregation. Fielding argues that open, decentralized systems will outscale walled-garden AI giants like OpenAI by harnessing global, incentivized contributions. The AI token powers the network, with fees funding protocol sustainability through buy-and-burn mechanics, while robust security and anti-manipulation design ensure integrity. The episode concludes with a forward-looking discussion on the future of AI, trust, and decentralized intelligence. Key takeaways include: 1) Machine learning needs horizontal scaling via decentralized infrastructure to overcome the limits of centralized, vertically scaled models; 2) Information markets enable bidirectional trading of knowledge, creating a live, self-correcting world model; 3) Decentralized AI thrives on programmatic trust and verifiable computation, not just prediction; 4) Tokenomics are designed for long-term sustainability through fee-based buy-and-burn, not inflationary rewards; 5) Security is prioritized through rigorous audits and machine learning-powered oracles; 6) Ethical concerns are addressed at the application layer, not the protocol level; 7) The future of AI is not about monopolistic control but open, incentivized collaboration across humans and machines; 8) Jensen’s vision is to build the internet of intelligence, where every device contributes to a global, self-improving AI ecosystem.