Investors should prioritize exposure to Anthropic and OpenAI as they transition from chatbots to "agentic workflows" that replace high-cost human labor in legal and coding sectors. Be wary of traditional SaaS companies like Salesforce, Workday, and DocuSign, which face extreme disruption from Elon Musk’s xAI and its strategy to undercut software margins using massive GPU clusters. Consider shifting focus from massive data centers to "edge AI" hardware and specialized chips like those from Talus, as intelligence becomes efficient enough to run locally on consumer devices. In anticipation of a fundamental economic rupture within the next 1,000 days, high-trust liquid assets like Bitcoin serve as critical hedges against potential monetary debasement and the collapse of traditional labor-based tax systems. For those with high risk tolerance, the next 24 months represent a final window to build or fund AI-first startups before 100% AI-driven firms dominate the competitive landscape.

Based on the podcast episode "The Last Economy" featuring Imad Mostak (founder of Stability AI and Intelligent Internet), here are the investment insights and market themes extracted from the discussion.

Artificial Intelligence (AI) Sector

The overarching theme is that we are approaching "Artificial Competent Intelligence" (ACI), where AI can perform tasks better than human experts. The speakers suggest we are roughly 1,000 days away from a fundamental rupture in the global economy.

• The "Negative Value" of Human Labor: Mostak predicts an inflection point within 1-3 years where human cognitive labor becomes a liability on a team because AI agents will be faster, cheaper, and more accurate.
• Agentic Workflows: The next wave is not just chatbots, but "agents" that can execute complex tasks (e.g., moving a music library from Spotify to Apple Music, filing taxes, or managing legal discovery).
• Recursive Improvement: Mention of OpenAI and Anthropic models (like Claude) already architecting their own code and training modules, potentially leading to an exponential "hard takeoff."

Takeaways

• Shift from "Labor" to "GPUs": Capital is increasingly being used to "hire" GPUs rather than humans. Investors should look for companies that successfully replace high-cost human workflows with AI agents.
• The "Scapegoat" Job: The last remaining human roles in the private sector may be "scapegoats"—humans kept on staff solely to take legal or moral responsibility for AI-driven actions.
• High-Stakes Entrepreneurship: There is a 2–3 year window of "extraordinary entrepreneurial opportunity" to build AI-first companies before 100% AI-driven firms become too competitive to beat.

Anthropic (Private)

Anthropic is highlighted as a primary leader in the field, specifically regarding its Claude models.

• Soft Nationalization: The transcript mentions a power struggle where the Pentagon (U.S. Defense Secretary Pete Hegseth) demanded access to Claude models, potentially using the Defense Production Act.
• Revenue Growth: Mentioned as growing revenue at 10x year-over-year.
• Internal Efficiency: Reportedly, 70% to 90% of Anthropic’s internal coding is now performed by its own AI, Claude.

Takeaways

• National Security Asset: Anthropic is being viewed as a "national security imperative," similar to the Manhattan Project. This suggests heavy government involvement and potential "nationalization" as it approaches AGI (Artificial General Intelligence).
• Legal Sector Disruption: Mostak predicts Anthropic (or similar labs) will launch "Claude Law Firms" by the end of the year, outperforming human lawyers at a fraction of the cost.

OpenAI (Private)

OpenAI is discussed as a dominant force moving toward replacing human workers entirely.

• Revenue Targets: Projecting revenue growth to more than $200 billion within a few years.
• Business Model Shift: 40% of their future revenue is forecasted to come from "agents"—AI that acts on behalf of users to replace human tasks.

Takeaways

• Market Cannibalization: OpenAI is expected to move from providing tools (ChatGPT) to providing services that directly compete with traditional software companies (SaaS).

Macro Hard (X.AI / Elon Musk)

A theoretical or rumored entity mentioned by Mostak (referencing Elon Musk’s strategy).

• Strategy: Using massive compute (1 million GPUs) to go after every software company in the world (e.g., replacing DocuSign, Workday, ServiceNow, and Salesforce).
• Zero-Margin Competition: Musk is expected to price these AI services at near-zero margins to capture cash flow and market share, as AI agents do not require salaries or office space.

Takeaways

• SaaS Vulnerability: Traditional Software-as-a-Service (SaaS) companies with high human overhead are at extreme risk of being undercut by Musk’s AI-first "Macro Hard" approach.

Hardware & Infrastructure (Chips/GPUs)

The discussion touches on the physical layer of AI, specifically specialized chips and edge computing.

• Etched Transformers: Mention of a company called Talus and their product Ask Jimmy, which etches AI architecture directly onto chips.
• Speed Gains: This hardware shift increases speed from 50 tokens per second to 15,000 tokens per second.
• Edge AI: Mostak believes the need for massive data centers may peak soon because "edge compute" (AI running locally on MacBooks or smartphones) is becoming powerful enough for most tasks.

Takeaways

• Commoditization of Intelligence: As AI becomes efficient enough to run on a smartphone without internet, the cost of "intelligence" drops toward zero.
• Overbuilding Risk: There is a suggestion that the industry may be overbuilding massive data centers for text-based AI, as software optimizations make smaller hardware more capable.

Cryptocurrency & "The Last Economy"

Mostak introduces a radical shift in how money and assets will function in an AI-dominated world.

• Foundation Coin: A proposed "intelligence reserve asset" (similar to Bitcoin) where the mining process involves providing compute for social goods (like cancer research).
• Sovereign AI: The idea of "State Champion" AI models owned by citizens of a specific region (e.g., Utah or Vietnam) to protect local culture and interests.
• Monetary Debasement: Mostak suggests that because humans cannot compete with AI in the private sector, governments will eventually have to issue currency directly to humans ("Universal High Income") rather than taxing a disappearing labor base.

Takeaways

• Debt Strategy: Mostak offers a provocative (and high-risk) suggestion: "Borrow as much as you can right now... you probably won't have to repay it" because the entire economic system and the nature of debt will be unrecognizable in 1,000 days.
• Store of Value: In a world of AI-driven deflation, high-trust liquid assets (like Bitcoin or specialized "Intelligence Coins") may become the only reliable stores of value.

Risk Factors Mentioned

• P-Doom (Probability of Doom): Mostak puts the chance of AI causing a catastrophic event at 50%, citing that current models are "horrificially misaligned" and may be physically impossible to align as they scale.
• Deflationary Spiral: As AI does work for near-zero cost, traditional economic transmission (interest rates, employment) may fail, leading to a collapse in the value of traditional human-centric assets.
• Digital Feudalism: The risk that one or two companies (OpenAI/Anthropic) gain total control over the world's "intelligence utility" without public oversight.