David Levine

David Levine is a technology entrepreneur and systems architect, the founder of Kiduna Club, and the founder, chairman, and president of the Kinship Intelligence Institute. His work spans early commercial web systems, financial marketplaces, multiplayer-game technology, and rooftop-solar modeling. With Kiduna, he is developing member-governed agentic organizations: communities whose software agents coordinate work, resources, and decisions around purposes defined by their members.

Early web ventures and the influence of LambdaMOO

Levine founded HuskyLabs in the early 1990s to develop public and enterprise web systems. He wrote Live Java: Database to Cyberspace, published by Academic Press in 1996, and subsequently founded ventures including Ultraprise, a marketplace for mortgage loans and loan portfolios, and Butterfly.net, a multiplayer-game technology company that became part of the business later known as Gamebryo. His historical professional biography traces these businesses alongside his later public-sector technology work.

An experience that continues to inform his thinking came in November 1993, when he joined LambdaMOO, the text-based virtual world hosted at Xerox PARC. Participants could write programs, adopt one another’s player classes, and transform objects within a coherent shared environment. In his talk on open-internet agent commerce, Levine recalls this composability as a source of community: people could extend the world together rather than simply consume what its operators supplied. He describes its governance, technology, funding, and culture as interdependent. His criticism of advertising-driven platforms follows from that experience: he argues that optimizing engagement and extracting value can displace the relationships that make a community worth joining.

From geospatial modeling to rooftop solar

After his game-technology venture, Levine worked in technology commercialization and geospatial modeling. He supported technology-based economic development in West Virginia and served as executive director of the National Technology Transfer Center. At Sewall, his work included natural-gas distribution modeling and wind-farm siting datasets; ImageTree introduced him to airborne LiDAR, and Lanworth brought him into satellite-based agricultural modeling. These experiences connected simulation and computing to decisions about physical infrastructure.

In 2010, he founded Geostellar, assembling former collaborators to apply computing, simulation, and remote sensing to rooftop solar. The task was to determine how sunlight reaches an individual roof and connect that potential to an installation and financing plan. In his account of Geostellar’s beginnings, he credits Breighton Dawe with the solar-radiance algorithms, Butch Deal with the computing infrastructure, and Jeremy Dobrzanski with the integrated application. Levine’s contribution was bringing those capabilities together in a solar marketplace, where the model could help a customer act on a roof’s energy potential.

He subsequently founded Indeco, pursuing solar financing through blockchain-based shared ownership. This period also included a business setback: Geostellar filed for bankruptcy in January 2018. His later work brings decentralized infrastructure together with questions about how organizations make decisions and distribute the benefits of their activity.

Kiduna and member-governed agentic organizations

With Kiduna, Levine applies those concerns to agentic AI. His founder account distinguishes three parts of the effort: Kiduna Club develops and stewards the supporting technology; Kinship Duna is the member-governed organization; and the Institute pursues research and public-interest applications. Within Kinship Duna, he describes himself as a catalyst assembling the foundations and initial participants, rather than an owner or executive above the membership.

• Agent identity: His proposal begins with cryptographic agent identity. Agents interacting across the open internet would use JWT credentials containing claims about their permissions, expiration times, and blockchain addresses. Other participants could validate those credentials before cooperating, connect the agent to a registered organization, and inspect an audit trail linking actions to the organization responsible for them. In his agent-commerce talk, Levine presents this as a response to the combination of private-data access, untrusted content, and external actions that makes prompt injection dangerous. The proposal explains how identity and delegated permissions could support accountability; it does not demonstrate that credentials or blockchain records eliminate prompt injection.
• Organizations as software: Levine also proposes organizations designed as software. An agent receives relevant knowledge, instructions defining its role, connected accounts and tools, specific abilities, and a purpose. Organizational software would then coordinate those agents’ work, including customer discovery, resource use, and reinvestment. He has chosen decentralized unincorporated nonprofit associations, or DUNAs, as the legal form for this model. His initial umbrella organization is intended to let builders affiliate their agents without first creating separate organizations, then develop organizations for their own purposes. In the talk, sales, social-media, and legal agents are examples of the templates he wants builders to contribute, rather than evidence of completed deployments.
• Decision markets: For governance, Levine proposes decision markets in which members introduce policies and participants trade pass and fail tokens. He argues that exposing participants to the consequences of their judgments could make deliberation more useful than persuasion alone. The intended role of agents is to bring members’ purposes, values, experiences, and aspirations into those competing judgments. This is a governance design proposal; the supplied evidence does not establish that it reliably produces better decisions than voting or other methods.
• Relational agency: His Genesis Nightpaper develops a related idea of relational agency: an agent should understand whose purpose it serves, where information came from, what permission applies, and who will experience the consequences of an action. The architecture separates memory and authority for people, relationships, projects, and organizations. A project could retain shared decisions without absorbing every participant’s private knowledge, while a tool’s ability to act would remain distinct from its permission to do so.

These distinctions give substance to Levine’s goal of expanding collective capacity while keeping control with participants. His solar work connected physical modeling to installation and financing; Kiduna seeks to connect agent capabilities to organizational purpose, permission, and governance. The Institute extends that agenda to public-interest applications, including ecological stewardship and community resilience, where commercial incentives may be limited.