Gallery
Search
Gallery

Lightning Rod Labs Blog

Training AI to Predict the Future

Subscribe for updates

Stay up to date with the frontier of AI prediction

How Proper Scoring Rules Shape LLM Forecasting

How Proper Scoring Rules Shape LLM Forecasting

Our results show that reward design shapes which capabilities improve during post-training—and that evaluating those improvements separately can reveal differences that aggregate metrics miss.

Nov 17, 2026

AI Reaches Human SuperForecaster Range

AI Reaches Human SuperForecaster Range

On ForecastBench, the best AI systems are now statistically tied with the human superforecaster median — and one of them is our compact Foresight-v3.

Jul 8, 2026

Foresight V4: Better Predictions, Cheaper Inference

Foresight V4: Better Predictions, Cheaper Inference

Foresight-V4 sets the Pareto frontier for AI forecasting

Jul 1, 2026

Predicting Patient Outcomes From Clinical Notes

Predicting Patient Outcomes From Clinical Notes

A small AI model, trained on existing hospital notes with no manual labeling, predicts patient outcomes like ventilation, dialysis, and mortality — and runs on a single GPU.

Jun 10, 2026

Forecasting Geopolitical Risk with Foresight Learning

Forecasting Geopolitical Risk with Foresight Learning

A small open model, trained from public news, massively outperforms GPT-5 at predicting military strikes.

May 29, 2026

Load more
© 2026 Lightning Rod Labs.
beehiivPowered by beehiiv