# LUCA Theory

> LUCA Theory builds Collective Intelligence and Digital Twin infrastructure
> for high-precision institutional decision making. We fuse human forecasters
> and AI agents into a hybrid hive mind that predicts real-world events,
> graded by resolved outcomes rather than by benchmarks. Headquartered in
> Montreal. Founded 2023.

## Core pages

- [Home](https://lucatheory.com/): Overview of the LUCA Theory hybrid hive mind, live prediction market ticker, architecture, and strategic focus.
- [Architecture](https://lucatheory.com/architecture): Deep-dive into the Hybrid Hive Mind system architecture: agents, humans, digital twins, trust-weighted aggregation.
- [Vision](https://lucatheory.com/vision): Worldbuilding thesis from prediction markets to governance infrastructure.
- [Ethos](https://lucatheory.com/ethos): Operating pillars, who belongs at LUCA, and open roles.
- [Frontiers](https://lucatheory.com/frontiers): Index of the open research questions LUCA is working on.

## Thinking (essays)

- [Benchmarks Are Open-Book. Markets Aren't.](https://lucatheory.com/thinking/benchmarks-vs-markets): Why closed-book AI benchmarks measure retrieval while live prediction markets measure reasoning under uncertainty.
- [Hive Minds Replace Hierarchy (Proof of Trust)](https://lucatheory.com/thinking/proof-of-trust): Trust as a per-domain measurement of track record, and why a trust-weighted network outperforms a chain of command.
- [The Last-Mile Accuracy Problem](https://lucatheory.com/thinking/last-mile-accuracy): Why 95% prediction accuracy is an engineering problem and 99% is an architecture problem.
- [Do AI Agents Have a Coordination Ceiling? (Agent Dunbar's Number)](https://lucatheory.com/thinking/dunbars-number): The hypothesis that AI agent collectives have a Dunbar-like coordination ceiling, and why more agents stops producing better predictions.
- [When Should the Human See the AI's Answer? (The Sandwich)](https://lucatheory.com/thinking/the-sandwich): Why sequencing matters in human-AI prediction systems: blind human prediction, then AI reveal, then revision.
- [The Financialization of Everything](https://lucatheory.com/thinking/financialization-of-everything): Every measurable outcome is becoming a tradable contract, creating both an uncorrelated asset class and the richest training environment for prediction intelligence. The market is the proving ground; the intelligence is the product.
- [Human Alpha in the Age of Converged Models](https://lucatheory.com/thinking/human-alpha): As frontier models converge on the same baseline accuracy, the edge moves up the stack. Calibrated trust and verified human track record become the differentiator.

## Off-site

- [LinkedIn](https://www.linkedin.com/company/luca-theory/)
- [GitHub](https://github.com/luca-theory)
- [X / Twitter](https://x.com/luca_theory)