The Machines Are Taking Our Jobs - Thank God? Emad Mostaque’s Guide to the next 1000 Days
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
This episode features Ahmad Mustak, founder of Stability AI and author of "The Last Economy," discussing his groundbreaking intelligence theory that applies generative AI mathematics to economics. (00:52) Ahmad argues that society is approaching a critical transition where AI capabilities will create massive abundance in intelligence while simultaneously disrupting traditional scarcity-based economic systems. The conversation explores the "abundance trap" - where post-scarcity technology is processed as poverty by current economic frameworks - and the "metabolic rift" between AI systems that don't consume like humans. (27:02)
- Main Theme: The need for new economic frameworks and social contracts to navigate the transition from scarcity-based to abundance-based economics as AI capabilities rapidly advance
Speakers
Nathan Labenz
Host of The Cognitive Revolution podcast, focusing on AI technology trends and implications. Nathan regularly interviews leading AI researchers, entrepreneurs, and thinkers to explore the transformative potential of artificial intelligence.
Ahmad Mustak
Founder of Stability AI (creator of Stable Diffusion) and current founder of the Intelligent Internet. Author of "The Last Economy: A Guide to the Age of Intelligent Economics." Ahmad is a signatory of the famous 2023 AI pause letter and has been instrumental in advancing open-source AI development. He's recognized as one of the most thoughtful voices in AI, combining technical expertise with deep consideration of societal implications.
Key Takeaways
The Intelligence Inversion Is Here
We're witnessing an unprecedented shift where intelligence - historically the domain of humans - is becoming abundant through AI. (21:58) Ahmad explains that by next year, anyone with a smartphone will have access to AI medical advice that outperforms human doctors, with 8 billion parameter models running locally that surpass human capabilities. This represents a fundamental inversion where we've moved from competing on land, then labor, then capital, to now intelligence itself. Unlike previous transitions where humans could pivot "up the stack," there's nowhere left to go - intelligence was our final competitive advantage.
The Abundance Trap Will Disguise Progress as Poverty
The most dangerous aspect of the AI transition isn't technological failure but economic misinterpretation. (27:07) Ahmad describes the "abundance trap" where genuine post-scarcity in intelligence will be processed by scarcity-based economic systems as widespread poverty and job loss. Even as AI dramatically improves human capabilities and access to expertise, traditional economic metrics like GDP will show decline due to displaced workers and collapsed pricing in knowledge work. This creates a perception crisis where objective improvements in human welfare appear as economic catastrophe.
Current Economic Models Are Breaking Down
The mathematical foundations underlying modern economics are becoming obsolete as AI violates core assumptions. (53:22) Classical economics assumed humans would remain the primary producers, but AI systems don't follow scarcity principles - they can be replicated infinitely, don't consume physical resources like housing or food, and continuously improve rather than degrade. Ahmad points to harbinger indicators like increasing recovery times from economic shocks, system fragility, and the growing disconnect between monetary value and actual human welfare as signs we're approaching a critical transition point.
A New Framework: Mind Capitals (M.I.N.D.)
Traditional economic measurement focusing solely on material capital (GDP) misses three other critical dimensions that determine success in an AI-abundant world. (62:14) Ahmad proposes the M.I.N.D. framework: Material capital (traditional wealth), Intelligence capital (capabilities and knowledge), Network capital (connections and trust relationships), and Diversity capital (optionality and adaptability). Success requires optimizing all four multiplicatively - if any reaches zero, the system fails. Countries like Singapore succeed by balancing all dimensions, while resource-cursed nations focus only on material capital.
Open Source AI as Democratic Infrastructure
The solution to preventing AI feudalism lies in treating AI models like public utilities rather than private assets. (82:05) Since AI training data represents the collective knowledge of humanity, Ahmad argues the resulting models should have collective ownership, especially for critical sectors like healthcare, education, and governance. He proposes building open-source models that are transparently trained and aligned with human flourishing, funded through a new cryptocurrency called FoundationCoin that channels investment directly into beneficial AI research and deployment.
Statistics & Facts
- 95% of corporate AI deployments have failed because companies are using models that are six months old, while AI capabilities are advancing exponentially. (17:45) This massive failure rate demonstrates the critical importance of staying current with AI developments.
- If every adult American received $16,000 annually (the poverty line) as universal basic income, the total cost would be $5 trillion - exactly equal to the entire U.S. tax base, making traditional UBI mathematically impossible. (38:06)
- Ahmad's medical AI model with 8 billion parameters outperforms human doctors and runs on just 8GB of RAM, costing approximately 2 cents in electricity for hours of operation on a laptop battery. (29:45)
