The AI Supercycle with Jordi Visser | Raoul Pal the Journey Man

Key Concepts

• Compute vs. Energy: The shift from a labor-capital economic model to one driven by the output of intelligence per unit of energy.
• Agentic Economy: The rise of autonomous AI agents that act as digital employees, shifting the Total Addressable Market (TAM) from humans to "infinity."
• Reed’s Law: The concept that the value of a network scales exponentially ($2^n$) as the number of nodes increases, now applied to AI’s self-recursive learning.
• Bottleneck Theory: The idea that supply-demand mismatches (chips, power, infrastructure) concentrate capital and drive innovation to clear these hurdles.
• The Vault: A personal, persistent memory layer where all user data (transcripts, documents, history) is stored and queryable by AI.
• Digestion/Consolidation: The necessary period of market cooling after a rapid "hockey stick" growth phase, allowing capital to rotate into new sectors.

---

1. The Shift to an Exponential Age

Raoul Pal and Jordi Vissa argue that we are witnessing a fundamental transition in the global economy. Traditional business cycles—driven by hiring, borrowing, and physical expansion—are being replaced by a cycle defined by compute and energy.

• The "Super Cycle": Because AI is self-recursive (improving its own algorithms), it creates a "double exponential" growth curve.
• The Bottleneck Reality: Shortages in chips and power are not signs of failing demand, but rather evidence of demand that is "too big." These bottlenecks force massive capital expenditure (capex) to solve for energy efficiency and infrastructure, which in turn accelerates the next phase of growth.

2. The Agentic Economy and Human Software

The speakers emphasize that the "Agentic Economy" is the most significant development of the current cycle.

• Digital Employees: Unlike human workers, AI agents do not consume traditional goods (housing, education). They consume compute. This changes the fundamental nature of economic growth.
• Human Software: Vissa highlights that companies like Eli Lilly (via GLP-1 drugs) are effectively "sucking in capital" to finance the next stage of human biological optimization, which he views as a form of "human software" development.

3. Methodologies for Navigating the AI Age

Both speakers share their personal frameworks for staying ahead:

• Knowledge Brains: Using tools like NotebookLM or RAG (Retrieval-Augmented Generation) vector databases to upload transcripts of expert thinkers (e.g., Jensen Huang, David Ricks). This allows users to "brainstorm" with the most intelligent minds in specific fields.
• The "Vault" Operating System: Vissa advocates for creating a personal, persistent memory layer. By centralizing all personal and professional data, an individual can use AI to instantly retrieve information, manage estates, or analyze historical patterns in their own life and work.
• Hardware Tinkering: Both emphasize the importance of "putting in the reps." They suggest that individuals must buy the latest hardware (e.g., high-end MacBooks, Nvidia setups) and actively use AI tools in daily life—walking, driving, or traveling—to build the intuition necessary to survive the transition.

4. Market Outlook and Rotations

• The Capex Buffet: The current infrastructure buildout is described as an "all-you-can-eat buffet." The recent wave of massive IPOs (e.g., Google, Cerebras) may signal a peak in the infrastructure capex trade, leading to a 3–6 month period of "digestion" or consolidation.
• Rotation to Laggards: As the "Magnificent 7" stocks digest their gains, capital is expected to rotate into the application layer, biotechnology, and eventually crypto.
• Crypto’s Role: Pal and Vissa argue that crypto is currently in a "digestion" phase. However, they maintain that the "third wave" of crypto is inevitable because the Agentic Economy will require financial guardrails, transparency, and tokenization to function at scale.

5. Notable Quotes

• Jordi Vissa: "The TAM [Total Addressable Market] was always humans, and now the TAM is infinity."
• Raoul Pal: "Intelligence is not software. Intelligence is intelligence. And it can have any output that you want or any input that you can give it."
• Jordi Vissa: "If you’re not learning how to use it [AI] now, I don’t know if you can catch up."

Synthesis and Conclusion

The conversation concludes that we are in a period of extreme, rapid change that is difficult for the human brain to process due to "bloated context windows"—the tendency to rely on historical data points (like the 1929 or 2000 crashes) that no longer apply to an exponential, AI-driven world. The key takeaway is that patience and active participation are required. While the market may experience "sloppy" periods of consolidation, the underlying secular trend of AI-driven productivity is unstoppable. Investors and individuals should focus on building their own "vaults" of knowledge, embracing the agentic shift, and preparing for a world where the cost of intelligence drops toward zero.