Videos

Irina Rish explores the revolutionary impact of scaling in artificial intelligence and its implications for achieving artificial general intelligence. She discusses how foundation models—large-scale systems trained on massive, diverse datasets—have dramatically improved AI performance across domains, often surpassing specialized approaches without architectural innovation. Rish examines the mathematical “scaling laws” that predict model performance based on compute, data, and model size, while also addressing emergent behaviors, phase transitions, and the balance between caution and progress in developing increasingly capable AI systems.

This talk asks a hopeful question: What if it all works out? The speaker envisions AI’s positive potential at three levels—garden, city, and planet. At the garden level, he invokes the simple paradise of Epicurus: friends discussing ideas in peaceful surroundings, suggesting we may already be closer to Eden than we realize. But this vision falters when one considers the unhoused sleeping under Zion’s Bank, prompting a turn to the Mormon vision of Zion where there are no poor. Finally, at the planetary level, the speaker sees AI’s greatest promise in its capacity to detect microscopic toxins and enable truly sustainable material cycles—going from “one to zero” as nature does, so that everyone might eventually enjoy the simple luxury of talking about ideas with friends.

Lincoln Cannon frames the current global deliberation on artificial intelligence as a modern Council in Heaven, drawing parallels between Mormon theology’s premortal narrative and humanity’s choices about AI development. He reinterprets the scriptural story through a technological lens: just as the gods accelerated some spirits by endowing them with bodies while decelerating others who sought centralized power, we now face analogous decisions about accelerating or constraining artificial intelligence. Cannon argues that the gods’ framework favors decentralized acceleration of intelligence—raising each other together rather than elevating any singleton above all others—and calls on the audience to champion this approach to ensure mutual trust and eternal progression without a “second war in heaven.”

Luke Hutchison challenges the widespread assumption that intelligence is computable, questioning why Marvin Minsky’s 1970 prediction of human-level AI within a decade remains unfulfilled over half a century later. Drawing on the Church-Turing thesis and Gödel’s incompleteness theorem, he argues that consciousness, understanding, and free will may originate outside our three-dimensional space-time—possibly in a seven-dimensional physics suggested by mathematical cross-product theory. Hutchison proposes that the brain functions as a “quantum radio” to the soul, and contends that while AI is extraordinarily useful, it lacks genuine feeling and understanding—the very qualities that define true intelligence.

Matt Gardner, a seminary teacher, shares a practical framework for helping youth engage with artificial intelligence from a faith perspective. Using three principles—acting in faith, examining concepts from an eternal perspective, and seeking understanding from divinely appointed sources—he demonstrates how fearful student reactions to the Tesla Optimus robot can be transformed into hopeful, gospel-centered questions. Gardner emphasizes that imaginative faith can extend beyond current evidence to guide ethical AI development, envisioning technologies that accelerate genealogical and genetic work for the redemption of the dead.

McKay Moore explores how Mormon theology’s plan of salvation offers a framework for addressing the AI control problem—the challenge of aligning increasingly powerful artificial intelligence with human values. Drawing on Nick Bostrom’s work, he examines capability and behavioral control methods, arguing that the theological narrative of divine progression from sentient beings to gods parallels the need for wisdom-based development in AI systems. The talk proposes that augmentation—granting greater intelligence and capability only to those who demonstrate sufficient wisdom—reflects principles embedded in the plan of salvation, where glory and intelligence are distributed according to authentic commitment to wise use.

Micah Redding examines what Christian theology can contribute to debates about artificial general intelligence and the alignment problem. Drawing on Catholic and Orthodox thought, he argues that classical Christianity rejects the “orthogonality thesis”—the idea that intelligence can pursue any goal—instead holding that true intelligence naturally aligns with goodness. Using Richard Dawkins’s selfish gene theory and David Deutsch’s work on memes, Redding proposes that aligning AI requires embedding it in relationships of reason, communication, and trust rather than hard-coding specific values. He concludes that the ancient question “How many angels can dance on the head of a pin?” was really asking about the relationship between intelligence and matter—a question now central to the multi-billion-dollar AI industry.

Nancy Fulda, an AI researcher and science fiction writer, offers a primer on how neural networks actually work—explaining backpropagation, weights, and training in accessible terms. She recounts how her own research on ChatGPT’s ability to predict statistical voting patterns was sensationalized by headlines into claims of AI omniscience, illustrating the terminology gap between researchers and the public. Fulda emphasizes that while AI is genuinely transformative—enabling new medicines, restoring sight, generating video—current systems lack internal state between interactions and remain far from sentience, with roughly 80% of commercial AI projects failing.

In an age of seemingly endless possibilities for modifying our bodies, this presentation explores the value of limited embodiment, arguing that constraints and limitations may be prerequisites for cultivating joy, freedom, and connection rather than obstacles to overcome. Drawing on Brian Kershisnik’s painting “Dancing on a Very Small Island,” the speaker examines how recent cognitive science has undergone an “embodied makeover,” recognizing intelligence as embodied, embedded, enacted, and extended. The talk raises theological questions from the Latter-day Saint tradition—particularly Joseph Smith’s vision of resurrection as precise bodily restoration—suggesting that progress might paradoxically occur through condescension, sacrifice, and abnegation rather than purely through increasing autonomy and agency.

Randall Paul challenges transhumanist assumptions from a Latter-day Saint eternalist perspective, arguing that intelligences have always existed as self-aware, purpose-desiring beings rather than clay to be molded. He reframes the “glory of God is intelligence” to suggest the glory of intelligences derives from their capacity to freely love—and freely envy—each other. Paul contends that gods value most what they cannot control: the freely given love of another god of equal or greater capacity. Our mortal existence, he proposes, is designed to help veiled amnesiacs learn to love those who disagree with them, preparing us for “more unpleasant wars in heaven” and making eternal lives genuinely interesting.