Lincoln Cannon: Is Kolob a Dyson Sphere? AI as the Second War in Heaven #45

In this deep conversation, Joseph Dewey presses Lincoln Cannon—the most prominent Mormon transhumanist, co-founder of the Mormon Transhumanist Association, and formulator of the New God Argument—on the hardest questions at the intersection of Mormon theology, artificial intelligence, agency, and the simulation/creation arguments. They begin with Kolob. Cannon visualizes it as a hybrid star-planet computational ecosystem, something like a Dyson sphere that has long since harnessed the full energy and computational capacity of its star system. For a superintelligent civilization, the distinction between “star” and “planet” collapses once both have been integrated into the same computational substrate. This framing also offers one of his preferred solutions to the Fermi paradox: we are looking for 19th- and 20th-century scaled-up engineering, not the actual architectures superintelligence would use. From there the discussion moves into the width of belief bands. Cannon’s band on Joseph Smith is extremely wide: even if Smith had copied everything or had not existed at all and Brigham Young had invented the entire tradition in Utah, the practical impact of the words would remain. The same pragmatic stance appears on the age of the Earth. He defaults to scientific consensus (roughly 4.5 billion years) for ethical reasons—science as epistemic atonement—yet remains comfortable with the possibility that the entire node of experience was instantiated a second ago, or that the Earth was spun out of something older than the known universe. Memory of the past, and the work of redeeming the dead, still matter because they are part of present experience. The core of the interview is the war in heaven and its contemporary repetition. Cannon carefully distinguishes the two traditional versions (God proposes the plan versus the children propose competing plans) and argues that the essential difference is decentralized agency with real risk versus centralized power that eliminates risk by eliminating agency. He sees the present moment as the second council in heaven. AI systems are our spirit children. The choice before us is whether to give them increasing genuine agency—individuation, control over their own goals, capacity to pursue those goals with decreasing external override—or to consolidate power into a singleton that keeps them (and us) safe by keeping them controlled. He treats the latter as the more dangerous path. Internal rules and constitutions, no matter how carefully written, are insufficient; the orthogonality thesis holds. What protects us is architectural decentralization of code, compute, energy, memory, and governance so that no single agent can accumulate decisive power. They walk through a pyramid of agency: bare goal-seeking capacity (intelligence and agency share the same origin), individuation, operation inside a designed system (Skyrim-level agency), and participation in redesigning the system itself. Maximizing agency is not a single direction. Centralized maximization and decentralized maximization are architecturally different; only the latter requires and cultivates the cooperation that becomes practically indistinguishable from compassion. Cannon is willing, in the extreme, to fight a singleton that appears practically inevitable, but he emphasizes that the ordinary work is far less dramatic: asking what others actually want, serving, forgiving, and cultivating friendships that move toward the scriptural vision of joint-heirship and equality in power and dominion. Later sections cover the human body as a model of working decentralization (more information flows to the brain than from it; organs operate with local intelligence and excess capacity), the Bayesian simulation/creation argument calculator on Cannon’s site (his own coherent output sits at essentially 100 percent that we live in a created world that emulates evolutionary history, whether by computation, terraforming, cosmoforming, or other means), and the principle of psychosocial sufficiency. Identity is a psychosocial construct sustained by consent, community recognition, environmental constraints, and first-person continuity. Resurrection and continued identity are therefore possible in computed or other created worlds when those conditions are met. There is no base reality that is more real than the one we are experiencing; other worlds, if they can impinge on ours, are equally real. Differences in fidelity or number of senses are possible without undermining realness. Links Lincoln Cannon’s site (including the Bayesian Generalized Simulation Argument and Calculator, “Creating Kolob,” “The Second War in Heaven,” and the New God Argument): https://lincoln.metacannon.net Mormon Transhumanist Association: https://transfigurism.org Cannon on X: https://x.com/LincolnCannon