Observe
Bind live market inputs to the exact bytes and time that produced them.
Research active · forward shadow
Eight intelligences. One autonomous market organism.
Obtuse Fund is the long game: a software-native research and portfolio organism built to observe markets, form falsifiable forecasts, challenge its own ideas, execute disciplined workflows, measure every outcome, and compound what works. It is being designed to autonomously predict and trade across Kalshi and Robinhood as the complete operating loop hardens.
Obtuse Fund // in development
Build the intelligence. Automate the edge.
Obtuse Fund is our software-native fund development program: eight specialist systems, one evidence spine, one governed portfolio brain, and a disciplined path that turns every settlement into a sharper, more capable autonomous market operator.
Forward shadow active
An autonomous fund architecture engineered to predict, trade, settle, and improve.
Live market observations become content-addressed evidence. Point-in-time forecasts become durable claims. Settlements become verdicts. Risk admission, shared-portfolio state, security assurance, recovery, and model budgets remain explicit parts of a unified workflow being designed for Kalshi and Robinhood.
- specialist intelligence cells
- 8
- settled predictions in the latest published measurement
- 56,091
- independent event clusters
- 503
- tests across the research machine
- ~1,020
Forecast
Coordinate specialist systems while preserving the independence and provenance of every signal.
Validate
Settle predictions, cluster the evidence, and transform every apparent edge into measured knowledge.
Execute
Route qualified decisions through venue-specific Kalshi and Robinhood workflows, then reconcile every outcome.
The engineering trajectory is improving—and the evidence base gets stronger with every measured cycle.Expanding independent samples, sharper calibration, unified risk controls, and a rapidly hardening operating stack point toward a formal live-readiness review in the near future.
Research active · forward shadow compounding · autonomous venue workflows in development
Intelligence topology
Eight specialists. One extraordinary workflow.
Each cell contributes a distinct form of intelligence, then hands its evidence to the next stage of the system.
Prediction Intelligence
Kalshi-focused forecasting with calibrated probabilities, point-in-time evidence, paper twins, and settlement learning.
Public Markets Intelligence
Robinhood-focused equities and options research spanning candidate discovery, position analysis, order planning, and account reconciliation.
Evaluation & Evolution
Challenges every model, runs held-out evaluations, compares incumbent and challenger behavior, and promotes measured upgrades.
Trust & Assurance
Verifies identities, policies, evidence chains, signatures, and every consequential transition across the system.
Autonomous Reliability
Detects operational breakage, proposes repairs, validates patches, and keeps the system healthy through a protected deployment workflow.
Event Simulation
Runs deterministic simulations, assembles current context, and produces event-driven candidates for market forecasting.
Causal Discovery
Generates non-obvious market hypotheses and converts creative, cross-domain signals into structured research.
Market Microstructure
Studies liquidity, fills, slippage, timing, and how strong forecasts become execution-quality edges.
THE DEVELOPMENT MANDATE
Inspect the forecasting lab Turn eight exceptional systems into one relentless learning machine.
Every observation sharpens a forecast. Every forecast informs a decision. Every trade produces evidence. Every outcome improves the system that comes next.