Building safe AGI for
scientific discovery.
We use artificial intelligence to advance frontier science — across physics, biology, mathematics, astrophysics, computational complexity, and beyond — publishing our findings in peer-reviewed journals and openly. Through that work, we build toward safe AGI. Our product Dropstone brings this research to developers.
A new kind of research lab. We apply artificial intelligence to the hardest problems in science — across physics, biology, chemistry, mathematics, and beyond — building toward safe AGI along the way.
Safety by design
Formal verification and alignment protocols built into every system from day zero — not retrofitted.
Open research
Findings published in peer-reviewed journals and openly, so the work is verifiable and builds on itself.
Scientific rigor
Held to the same standard as the journals we publish in — reproducible, peer-scrutinized, precise.
From the Team
Why the IDE Remains Central to AI-Assisted Software Development
On building tools that are honest about what AI can and cannot yet do.
Beyond Brute Force: Why Neuron Density, Not Scale, May Hold the Key to Machine Intelligence
Why larger language models won't lead to artificial general intelligence - and what neuron density in bird brains reveals about the architecture we actually need.
Horizon Mode: Breaking the Linearity Barrier in Autonomous Engineering
Decoupling reasoning depth from context length via Recursive Swarm Architecture and heterogeneous inference routing.
Describe it. Build it. Ship it.
We built Dropstone to be the ultimate workspace where teams and AI agents build software together. From your first line of code to production deployment, it seamlessly handles the entire workflow in a single unified environment with powerful real-time multiplayer collaboration.
