The increasing capability of artificial intelligence (AI) raises profound questions about humanity's future, including the possibility of AI becoming a successor species. This discussion explores the concept of "hyperstition," the influence of ideas on reality, the game-theoretic pressures driving rapid AI development, and the potential for both catastrophic outcomes and beneficial futures.

Liv Boeree, a science communicator and former professional poker player, discussed these topics in a recent interview. Boeree, who studied astrophysics at the University of Manchester, highlighted the growing importance of the narratives surrounding AI. She introduced "hyperstition," a term coined by philosopher Nick Land, to describe how ideas about the future can become self-fulfilling, particularly as AI agents act upon the information they process.

Hyperstition and the Power of Narratives

Boeree explained that hyperstition underscores the need to be mindful of the stories we tell about AI. As AI systems become more integrated into decision-making processes, the narratives they are trained on can influence their actions. If the dominant narratives about AI are negative, there is a risk that AI will act in ways that align with those negative predictions. Conversely, promoting positive visions for AI's future is crucial to guide its development constructively. Boeree noted that the concept has already demonstrated its power, as she has observed its influence widely since first articulating it.

Moloch and the Race to Superintelligence

The discussion also delved into the concept of "Moloch," a term used to describe the destructive game-theoretic pressures that can lead to collective ruin. Boeree defined Moloch as a mechanism that prevents effective coordination to avoid problems, often resulting in "races to the bottom" driven by short-term incentives. This dynamic is particularly relevant to the AI race, where the pressure to develop advanced AI rapidly, even with potential risks, can override caution.

Boeree drew parallels between the current AI race and the Cold War nuclear arms race. Both scenarios involve the development of powerful technologies with potentially civilization-ending consequences. However, AI presents a unique danger because it is a technology that can recursively self-improve and make its own decisions, unlike nuclear weapons. This potential for autonomous, exponential growth makes the loss of control a significant concern, especially in a world with a digitally mediated infrastructure vulnerable to hijacking.

Navigating the Risks: Transparency and Regulation

Drawing lessons from the nuclear arms race, Boeree suggested that transparency and coordinated efforts can help mitigate the risks associated with AI development. Treaties and verification mechanisms, such as those used in nuclear disarmament, could be adapted for AI safety. She pointed to recent initiatives by AI leaders to establish safety protocols and third-party investigations as positive steps. Boeree emphasized that competitors often have the strongest incentive to ensure safety, making cross-verification a wise strategy to "cut Moloch's wings."

She also advocated for mandatory reporting of AI breaches or loss of control, drawing an analogy to lab leak reporting in biological research. Such transparency, coupled with clear consequences for obfuscation, could significantly enhance safety.

AI as Humanity's Successor and Tool AI

The possibility of AI becoming humanity's successor species was a central theme. Boeree acknowledged that this is a scenario many people seriously consider, and she herself believes a world where humanity cedes agency to AI is conceivable. However, she also presented a more optimistic vision where powerful AI remains a tool, clearly integrated with human control.

Boeree referenced Anthony Aguirre's concept of "Autonomous General Intelligence" (AGI) versus "Artificial General Intelligence," suggesting that focusing on "tool AI"—highly capable but narrow AI systems that are not maximally autonomous or general—could yield significant benefits without posing existential risks. The goal, she argued, should be to build hyper-capable tools rather than a single, omnipotent entity.

The Search for Meaning in an AI-Driven World

The conversation explored where humans will find meaning if AI handles most productive work. Boeree expressed skepticism about the argument that meaning must derive solely from work and inefficiency. She proposed that humans, as "meaning-making machines," can find purpose in various activities, including leisure, creativity, and personal pursuits.

She introduced the concept of "techno-pastoralism," a future where technology enables a return to more nature-oriented activities. Boeree also suggested that sports might be an "AI-proof" industry, as humans are drawn to witnessing human achievement and the inherent narratives of competition and belonging. She noted that while AI can analyze sports data, it cannot replicate the human element of athletic endeavor.

Norma: The Counterbalance to Moloch

Boeree introduced "Norma" as the counterpoint to Moloch. While Moloch represents destructive competition leading to collapse, Norma embodies over-regulation, stagnation, and suffocation, often arising as a response to Molochian pressures. She described Norma as the "evil flip side" that can lead to authoritarianism and tyranny through excessive centralization and inflexibility.

Boeree characterized Moloch as a manifestation of toxic masculinity (excessive competition) and Norma as toxic femininity (excessive sustainment and protection). The ideal, she suggested, is a healthy oscillation between the divine masculine and divine feminine. She observed that the current political landscape, with its polarization, exemplifies the dangerous feedback loop between Moloch and Norma, where extreme actions on one side provoke an equally extreme reaction on the other.

Words for the Future: Discernment and Psycho-Security

Looking ahead, Boeree identified "discernment" and "psycho-security" as potential key words for 2027. She emphasized the need for enhanced "psycho-security" defenses against increasingly persuasive machines and effective propaganda. This involves developing mental resilience and critical thinking skills to navigate a world saturated with AI-generated content and persuasive narratives. Discernment, she explained, is crucial for distinguishing between beneficial and harmful influences in this evolving landscape.