Post-AGI economics for rational optimists
Many have written and talked about a future where work would be entirely done by AI. I'm rational in the fact that I don't believe work will disappear, but I'm optimist in that it will create new economic laws.
Many debate when AGI will arrive, but this misses the point. We're already witnessing the early stages of this economic transformation. GPT-4 isn't "full AGI," but it's already capable enough to reshape entire industries. The revolution isn't coming - it's here, just unevenly distributed and not yet operating at full scale.
One thing I'm almost sure - as AI eliminates operational friction, we're moving from a world of normal distributions to one of extreme power laws. And we shouldn't stay in the middle.
The Platform Hierarchy: How AGI Reshapes Business
Meta-Platforms: The New Utility Companies
The AGI economy structures itself like a neural network - with foundational layers (meta-platforms) and specialized layers (vertical platforms) built on top.
Think of OpenAI, Anthropic, and DeepMind not as tech companies, but as utility providers - the new electricity companies of cognitive power. Instead of paying salaries to knowledge workers, businesses will pay subscription fees to these meta-platforms for basic cognitive capabilities. And we're already doing it, many companies are paying LLM subscriptions to their employees.
Vertical Platforms: The Industry Specialists
Built on meta-platforms, vertical platforms replace entire professional categories. A legal vertical platform might cost $50,000 monthly but replace $500,000 in lawyer salaries.
These aren't just cheaper alternatives - they're exponentially more capable as they are specifically trained on datasets serving specific use cases. A single healthcare vertical platform could handle the cognitive workload of an entire hospital's administrative staff.
Key vertical domains include:
- Healthcare AGI: Replacing medical administrators, analysts, and support staff
- Legal AGI: Substituting for junior lawyers, paralegals, and compliance teams
- Financial AGI: Taking over from analysts, traders, and risk managers
- Scientific AGI: Replacing research assistants, data analysts, and lab managers
The New Economic Reality
This creates a three-layer economy where businesses pay for cognitive capacity rather than headcount:
- Meta-Platform Subscriptions: Base intelligence costs (3-5 global providers)
- Vertical Platform Fees: Industry-specific capabilities (dozens per sector)
- Business Operations: Organizations consuming these services (millions)
The genius isn't just technical - it's economic. A business might pay $100,000 monthly for AGI services that replace $1M in salary costs. But more importantly, these platforms can scale instantly, work 24/7, and perform at superhuman levels. The cost-to-capability ratio becomes exponential rather than linear.
As Taleb might observe, this structure creates both efficiency and resilience - businesses can instantly scale cognitive capabilities up or down, while maintaining human oversight for critical decisions. The meta-platforms provide the stable foundation, while vertical platforms offer the specialized tools that make human-AGI businesses possible.
The Zero-Marginal-Cost Platform Economy
The New Cost Structure
Traditional businesses balanced capital expenses against operational costs. The platform hierarchy changes this fundamentally. Meta-platforms bear massive fixed costs (compute, R&D) but offer near-zero marginal costs for cognitive operations. Vertical platforms inherit this economics - they have development costs but negligible per-use costs. For businesses consuming these services, information processing becomes a utility cost rather than an operational expense.
The Capital Concentration Reality
This creates three levels of capital concentration:
- Meta-platforms: Require billions in compute and R&D (OpenAI, Anthropic)
- Vertical platforms: Need significant specialized development capital
- Businesses: Shift from operational costs to platform subscription fees
The Platform Barbell
The economy splits into three distinct layers:
- Platform Providers: Meta and vertical platforms capturing most value
- Platform Orchestrators: Businesses combining and deploying platforms
- Human Specialists: High-touch services platforms can't replicate
New Platform Economics
The Network Effect Amplifier
Both meta and vertical platforms exhibit unprecedented network effects. Each user improves the system while adding virtually no cost. This creates a compounding advantage - the more a platform is used, the better and cheaper it becomes. Traditional diminishing returns are replaced by exponential improvements.
The Human-Platform Symbiosis
Value shifts to two extremes: platform orchestration and genuine human expertise. The winners will be those who can either:
- Master platform combination and deployment
- Provide human judgment that platforms enhance but can't replace
- Build or own pieces of the platform hierarchy
Strategic Imperatives
The Platform Imperative
Success requires positioning within the platform hierarchy:
- Build/Own platform components where possible
- Master platform orchestration and combination
- Develop platform-enhanced human capabilities
- Shift from operational excellence to platform leverage
Antifragile Platform Strategy
Organizations must:
- Diversify platform dependencies
- Maintain critical human capabilities
- Build unique platform combinations
- Focus on platform-native business models
The Path Forward
The rational optimist sees this platform hierarchy as liberation. When meta and vertical platforms handle cognitive heavy lifting, organizations can focus on unique value creation through platform orchestration and human expertise. The economy becomes an interconnected web of platforms and human specialists, each focusing on their comparative advantages.