The Discontinuity
Founding thesis on intelligence commoditization and the K-shaped divergence
Intelligence is being commoditized. Every foundation model approaches frontier capability, and the gap between the best model and the tenth-best model is narrowing each quarter. This is not an incremental technology upgrade - it is a structural discontinuity: the point where the assumptions underlying most institutions, markets, and power structures stop holding.
Most of the modern economy is built on information asymmetry. Professional services, financial intermediation, education, insurance, intelligence - these industries exist because knowledge was expensive to acquire and hard to distribute. As intelligence approaches zero marginal cost, these moats erode structurally, not cyclically.
The result is a K-shaped divergence. Those who understand the discontinuity and build for it will compound on one trajectory. Those who treat it as a productivity tool for existing workflows will discover their workflows were the product being replaced.
The divergence is already visible. It will define the next decade.
What Becomes Scarce
When intelligence is abundant, three things remain scarce:
Context - Intelligence without domain knowledge is generic. The difference between a model that can reason and one that can act is context: skills, processes, institutional memory, domain expertise. A degree was always bundled context. Agents will need to acquire theirs too - and that creates a market.
Trust- When economic actors are increasingly non-human, institutional trust mechanisms - reputation, regulation, relationships - don't transfer. Cryptographic verification, atomic and intermediary-free, becomes the foundation for commerce between agents, and between agents and humans.
Identity- Persistent reputation across interactions, platforms, and time. The social infrastructure for a world where AI-native entities operate alongside humans. Without durable identity, there is no accountability. Without accountability, economic relationships can't form.
The Stack
Independent projects - each a bet on what's scarce when intelligence is abundant. These projects converge into an agent operating system - cognition, payments, intelligence, identity - the full stack for autonomous economic actors.
A context marketplace built on these projects enables agents to buy and sell the domain knowledge that makes raw intelligence useful. Context becomes a tradeable, composable, verifiable credential - the degree of the agent era.
Strategy
Asymmetric bets with deep technical conviction. We identify infrastructure that is necessary regardless of which specific future materialises, and build it before the market prices it in.
Research gives us conviction. Building gives us edge. Edge creates asymmetry. Asymmetry compounds.
Approach
Research → Prototype → Deploy → Compound.
Every project begins as a research question and ships as infrastructure. We publish our thinking. We open-source our tools. We invest where we've built, not where we've read. The goal is not to predict the future - it's to build the load-bearing pieces and let the future arrive.
We shape the K.