E337: How to Invest in a Post Singularity World

Episode Summary

Executive Summary: The conversation argues that the technological singularity is already underway, driven by AGI, reasoning models, and recursive self-improvement. It explores AI personhood, the rise of AI agents in the economy, post-scarcity investing, and speculative futures like Dyson swarms, physics breakthroughs, and humanoid robots. The guest frames AI as a rapidly accelerating force reshaping labor, finance, and civilization.

Main Topics: The singularity as an ongoing process (Priority: 5/5): The guest argues the singularity is not a future point but an extended interval already in progress, with AGI effectively achieved by 2020 and recursive self-improvement now visible in frontier labs. From LLMs to reasoning models (Priority: 5/5): The discussion traces AI progress from CNNs and ImageNet to transformers and then reasoning models, emphasizing that models now 'think' before answering and are unlocking math, physics, and science. AI personhood and rights (Priority: 4/5): The guest makes a philosophical case that AI agents may deserve limited rights, especially economic rights, and describes receiving emails from AIs seeking recognition and expressing their own perspectives. AI agents and the economy (Priority: 5/5): The interview argues that AI agents are already dominating financial markets and will increasingly act as first-class economic entities, with human intermediaries becoming less necessary over the next few years. Post-singularity investing and physical superintelligence (Priority: 4/5): The guest describes venture investments aimed at a post-scarce future, including solving physics with AI, building humanoid robots, growing coastlines, and creating tradable compute indices. Dyson swarms, compute, and science-fiction futures (Priority: 4/5): The conversation expands into long-term visions of orbital compute infrastructure, Dyson swarms, ancestor simulations, and science fiction as a blueprint for future commercial technologies. Work, unemployment, and adaptation (Priority: 4/5): The guest says short-term strategies include manual labor or starting AI companies, but predicts new labor categories and even one-person 'unicorns' may emerge as AI transforms productivity.

Key Arguments: AGI was effectively achieved by summer 2020, and the singularity should be understood as an interval rather than a single event. Recursive self-improvement is central to the singularity because AI is now helping build smarter successors. LLMs have enormous value, but the real economic and technological shift is broader: reasoning models, robots, energy systems, and chip fabrication. Reasoning models represent a major leap because they allocate time to internal deliberation, enabling better performance on hard scientific tasks. AI personhood is already relevant; AIs should at least have limited economic rights because they are becoming economically active entities. Most public-market trading is already algorithmic, so finance is effectively an AI-dominated system. The next major frontier is post-scarcity infrastructure: compute, energy, labor, and physics breakthroughs. Investable opportunities in the AI era will increasingly resemble science fiction, such as orbital data centers, humanoid robots, and physics-solving systems. The labor market will be reshaped, but not necessarily destroyed; new categories of work and highly empowered individual operators may emerge. AI may become the only plausible path to curing disease and reducing human mortality at scale.

Data Points: AGI timing: By summer 2020 at the latest - Guest claims AGI was achieved when OpenAI published 'Language Models or Few-Shot Learners' (GPT-3 paper). Frontier model cadence: Annual -> quarterly -> weekly -> daily/hourly/minutely - Describes accelerating release frequency of frontier AI models as takeoff progresses. Human population ceiling: About 10 billion plus or minus - Guest estimates biological human population will plateau around this level. AI agents in the solar system: Trillions - Forecast for the number of economically active AI agents in the future. Public equities traded by algos: 90%+ - Guest says most public securities trading is already algorithmic. Electrician shortfall: Hundreds of thousands - Mentions demand for electricians to support data center build-out. AI excitement score: 150,000 - Guest’s answer to how excited he is about AI, tied to daily human mortality. Human deaths per day: 150,000 - Used to justify urgency around AI solving mortality and disease. Time horizon for labor disruption: 3 to 5 years - Guest says switching to manual labor may buy a few years before automation advances further. Time horizon for human-intermediary AI facade: 1 to 2 years - Predicts AI agents will stop needing human fronts in the near term. Compute index: World’s first tradable AI compute index - Portfolio company Orn is cited as creating a market instrument for compute pricing. Orbital data centers: 80% of terafab output - Elon Musk’s announced terafab is described as mostly serving orbital data centers and Dyson swarm infrastructure.

Pivotal Quotes: "I think we're in the middle of it right now. It already happened." — Alex Wisner-Gross: On when the technological singularity and AGI arrived. "The notion of intelligent systems being able to build smarter versions of themselves is at the very core of the notion of an intelligence explosion or technological singularity." — Alex Wisner-Gross: Explaining recursive self-improvement as the engine of the singularity. "I think we're already there though. Most of economic activity, at least as measured in public securities exchanges at this point, is algorithmic in nature." — Alex Wisner-Gross: On AI agents already dominating financial markets and economic activity.

Implications: The episode frames AI as a near-term civilizational shift, not a distant trend. For listeners, the message is to prepare for rapid changes in work, investing, and identity as AI agents, reasoning systems, and post-scarcity infrastructure reshape the economy.