We work on some of the most complex and interesting challenges in AI
From autonomous scientific discovery to foundation models trained on molecular structures, our research pushes the boundaries of what AI systems can reason about, discover, and create.
Breakthroughs
Publications
Books · IV Volumes
Papers
Latest News
From the Lab
Behind-the-scenes stories, technical deep dives, and insights from our research team.
Training Aethel: Building a Foundation Model for Science
The architecture decisions, data curation strategies, and training dynamics behind Aethel — our 1 trillion parameter scientific reasoning model.
Morbius' First Autonomous Discovery
How our agentic system independently formulated a novel hypothesis about catalyst behavior and designed the experiments to validate it.
Building Prometheus: An Agentic Research Assistant
The technical challenges of creating a long-horizon reasoning system that reads, synthesizes, and generates research insights from millions of papers.
Rethinking Portfolio Theory with Causal Inference
Why traditional mean-variance optimization fails under regime shifts, and how Parallax uses causal graphs to build robust strategies.

A Geometric Analysis of Quantum-Inspired Local Tensor Regression











