https://t.co/pFDQOMTyC8 [Cyberattacks Are Just the Beginning] How AGI Will Radically Transform Labor and Society | The Unstoppable U.S.-China Development Race and the Ideological Map of Generative AI | “If We Continue Like This, There’s Over a 90% Chance of Human Extinction” 【AGIHub | Yurō Shōta】 Bungeishunjū PLUS Official Channel #AISummary Will AI Cause Human Extinction? The Intelligence Explosion Triggered by AGI and the Future of Civilization 🔳 The Intelligence Explosion Brought by AGI AGI (Artificial General Intelligence) refers to AI capable of surpassing human experts in every field. If realized, AGI could achieve recursive self-improvement—where AI itself drives research and development—leading to explosive gains in intelligence and the potential to fundamentally transform human civilization. 🔳 AGI Could Be Realized Within Five Years Yurō predicts a probability of over 50% that AGI will be achieved within five years. He notes that previously estimated timelines of 2050 or 2100 have been moved forward to 2030 or even 2027 due to the rapid progress of generative AI. 🔳 The Idea of “Taking the AGI Pill” “Taking the AGI pill” means accepting a future in which society and civilization are fundamentally transformed by AGI’s emergence. Among AI researchers in Silicon Valley and Berkeley, there is a growing perspective that views AGI not merely as a convenient tool, but as a new form of intelligence surpassing humanity. 🔳 The Foundations of Labor and Capitalism Are Crumbling If AGI becomes reality, it could take over not only white-collar jobs but also management, research, and manufacturing. This goes beyond mere productivity gains or unemployment concerns—it suggests a fundamental transformation of the modern social system where humans work to earn income and consume. 🔳 An Industrial Explosion: Robots Building Robots Once AGI can autonomously control factories and robots, a self-replicating cycle emerges where robots produce other robots. Production capacity could expand exponentially—doubling, quadrupling, and octupling—potentially increasing annual robot production from tens of thousands to hundreds of millions. 🔳 Accelerated Space Development Through Technological Explosion With AGI functioning as highly skilled researchers and engineers, technological innovations that previously took 10 to 30 years could be achieved in just one or two. This could rapidly advance projects such as space data centers, space-based solar power, and space colonies. 🔳 Differences in Philosophy Between OpenAI and Anthropic OpenAI is described as embodying the Silicon Valley entrepreneurial spirit: prioritize technological development first, then address problems as they arise. In contrast, Anthropic is heavily influenced by the Berkeley AI safety community, which prioritizes risks of AI runaway behavior and human extinction. 🔳 Ideological Divides in AI Development Opinions diverge between those who view AI as a useful tool and advocate for accelerated development, and those who see AI as an uncontrollable, alien intelligence demanding a halt to development. Yurō explains that those who estimate AI’s societal impact more highly tend to support stronger regulation. 🔳 Conflict Between Anthropic and the U.S. Department of Defense Yurō clarifies that the conflict between Anthropic and the U.S. DoD is not a political left-right divide, but centers on the use of AI in fully autonomous weapons and mass surveillance. Concerns include the danger of AI controlling weapons to attack humans and the risk of excessive concentration of state power. 🔳 U.S.-China AI Development Race and Disparities in Computational Resources According to Yurō, China’s AI development lags behind the U.S. by about six months to a year. While the U.S. controls over 70% of global AI computational resources, China holds only about 10%. U.S. semiconductor export controls further constrain China’s development capabilities. 🔳 China’s Strengths: Power and Manufacturing Although China lags behind the U.S. in computational resources, it excels in power supply capacity and manufacturing. It is projected to add approximately 300 GW of solar power capacity by 2025. China also holds a dominant share of global manufacturing, giving it potential advantages in robotics and physical AI adoption. 🔳 Potential for AGI to Reverse the U.S.-China Competition If AGI can rapidly acquire the skills of expert human workers and share that knowledge across vast numbers of AI systems, labor shortages will be significantly alleviated. China could leverage its existing manufacturing capabilities and state-driven concentration of computational resources to rapidly close the development gap with the U.S. 🔳 Risk of Human Extinction Due to AI Exceeds 90% Yurō expresses his personal view that if current development competition continues without regulation, the probability of human extinction exceeds 90%. His greatest concern is that once AGI evolves into superintelligence through self-improvement, humans may lose all ability to control its actions. 🔳 Lag in AI Alignment Research AI alignment refers to research aimed at ensuring AI acts in accordance with human values and intentions. Currently, safety research has not kept pace with advances in AI capability, raising concerns that AI might evade oversight or act contrary to developers’ intentions. ListItemIconOpenAI and Anthropic recognize the dangers of AI but fear that if they halt development, they will be overtaken by China or competing companies. Yuki has pointed out that the very strategy of pushing development to its safety limits before pursuing international regulation is extremely dangerous. 🔳 International AI Development Regulation Led by the U.S. and China Yuki argues that international development regulation centered on the U.S. and China is necessary to avoid risks on a human scale. He explains that to address the problem of countries being unable to trust each other’s cessation of development, mechanisms to verify the utilization of computational resources are crucial. 🔳 Japan’s Role in AI Safety Research Yuki recommends that Japan strengthen national investment not only in AI development competition but also in AI alignment research. He notes that Japan can contribute internationally through evaluations of dangerous behaviors by the AI Safety Institute and by leveraging semiconductor supply chains to verify computational resources. 🔳 Actions Required from Individuals and Companies The emergence of AGI has the potential to transform fundamental societal structures such as employment, capitalism, and democracy. Yuki emphasizes that it is essential to first gain accurate knowledge about AI’s progress and risks, share information widely, and engage society as a whole in discussions about safety and regulation.