Crene is investment thesis review infrastructure. The platform decomposes macro, labor, AI transition, and investment thesis questions into falsifiable components using four frontier AI models (Claude, GPT, Gemini, Grok). The system produces three product types: thesis maps (binary decomposition into assumptions), factors (continuous decomposition into drivers), and scenarios (hybrid decomposition into coherent pathways). Cross model disagreement is the primary signal. Calibrated against 1,161 resolved events with a Brier score of 0.114.

How does AI prediction calibration work at Crene?

Calibration applies at the consensus layer, not at the underlying model layer. Crene queries four frontier models in isolation, computes a cross model median consensus, and tracks resolution against a tiered source allowlist. Every resolved event becomes a permanent calibration point. Per tier accuracy, per model Brier scores, and per domain accuracy are computed live and published at https://crene.com/methodology.

What is a Brier score?

A Brier score is the proper scoring rule for probability forecasts on binary outcomes. It is the mean squared error between the forecast probability and the realized outcome (0 or 1). Lower scores indicate better calibration. The no skill baseline is 0.25 (always predicting 0.5); a perfect forecaster scores 0.0. Crene publishes the consensus Brier score and per model Brier scores across the resolved event corpus.

Crene serves investment teams that need a repeatable thesis review process: hedge funds, asset managers, macro PMs, CIO offices, family offices, and thematic investment teams. A team brings one thesis it is actively debating. Crene maps what the thesis depends on, where models disagree, what changed each week, and what would force a rethink. Data and API access remain available as the proof layer behind the workflow.

What is the difference between an event, a thesis map, and a factor at Crene?

Crene has four product types. Events are base binary outcomes with scalar probabilities. Thesis maps decompose binary anchor questions into child situations. Factor maps decompose continuous variables into drivers. Scenarios map long horizon investment questions into coherent pathways across live systems including Empire by Default, AI Labor Transition, European Rearmament, and India as the Third Growth Pole.

Bring your own thesis

Get this same map built on a thesis your team is debating.

Assumptions, model disagreement, and rethink triggers, updated weekly before your PM, risk, or IC discussion. Currently accepting one macro thesis and one AI-economy thesis for July.

Map one thesis

Back to thesis map Part ofWill AI generated content exceed human generated content on the US internet before 2030?

Event · SOCIAL TRUST & INSTITUTIONAL RISK

Will US children under 18 consume more AI generated than human generated educational content by 2029?

Resolves Dec 31, 2029

Probability

40%

4-model average

Confidence

LOW

building category history

Stability

—

Models

Diverging

38pt spread

Current read

Consensus sits at 40% across the four models and is — (loading). Models are diverging at a 38pt spread, so the average conceals live disagreement worth inspecting below. Category confidence is LOW (building category history).

Confidence reflects category-level track record. Stability tracks estimate movement. Models shows whether the four agree.

Belief trajectory

Loading belief trajectory...

Per-model probabilities

GPT-4o

60%+20 vs avg

By 2029, the trend toward the integration of AI in education is expected to accelerate significantly, driven by advancements in AI technology and increasing demand for personalized learning experiences. For instance, a 2023 survey by Deloitte indicated that 65% of educators believe AI will play a critical role in the future of education, suggesting a growing acceptance and reliance on AI-generated content. Additionally, current digital learning platforms report that AI-generated materials are projected to grow by 25% annually, indicating a substantial shift toward AI content consumption among children.

Key uncertaintyThe regulatory environment concerning AI in education, particularly related to data privacy and content quality standards, could significantly impact the rate of adoption and acceptance of AI-generated content.

Gemini

45%+5 vs avg

The rapid advancement and increasing accessibility of AI-generated educational tools, coupled with a growing awareness of their potential benefits like personalized learning, suggest a significant rise in consumption. However, widespread adoption by 2029 hinges on overcoming challenges in content quality, educator trust, and equitable access, which may limit AI's dominance over established human-created resources. The current trajectory indicates a strong increase, but a clear majority is not yet guaranteed.

Key uncertaintyThe speed and effectiveness of AI content moderation and quality assurance mechanisms will be a crucial factor in building trust and encouraging widespread adoption.

Claude

22%-18 vs avg

Current AI-generated educational content represents less than 5% of US K-12 educational material consumption (based on 2024 market data showing AI edtech at ~$2.2B of $180B+ total US edtech spending), requiring a 20x acceleration in adoption and displacement of human-created content in 5 years. Historical technology adoption curves show that even rapidly-adopted edtech (like online learning, which took 15+ years to reach 30% penetration post-2005) move slower than this threshold requires. The definition's ambiguity around "more than" (>50% or just plurality) and measurement difficulty create additional headwinds, while structural factors like teacher resistance, curriculum standardization, and regulatory scrutiny around AI in education suggest adoption will likely reach 15-30% by 2029 rather than crossing the 50%+ threshold.

Key uncertaintyWhether large language model improvements and declining AI content production costs lead to institutional wholesale adoption by school districts (potentially accelerating to 40%+ penetration by 2029) versus continued fragmented, supplemental classroom integration.

Grok

34%-6 vs avg

US K-12 students currently spend ~6.5 hours/day on digital devices, yet only 12-15% of assigned educational content is AI-generated per 2024 NCES and Common Sense Media data; Khan Academy's GPT-4 tutor and Duolingo Max reached 4M+ US students in 2024, but state textbook adoption cycles average 7-8 years and 38 states still require human-authored core materials under existing procurement rules. Major districts (NYC, LAUSD) have expanded pilots to 8-12% AI content for 2025-26, while teacher-union pushback and FERPA compliance reviews constrain rapid scaling. Structural acceleration would require either federal ESSER-style funding or state textbook mandates shifting before 2028.

Key uncertaintyWhether state legislatures amend textbook procurement statutes to treat AI-generated core instructional materials as equivalent to human-authored texts before the next adoption cycle.

Key disagreementGPT-4o (60%) vs Claude (22%): Different weighting of factors

CRENE-AICD-SO10Generated Jul 3, 2026