Orion Archer Cohen
I am a computational chemist, compulsive generalist, and co-founder of Riven Systems. I am interested in how exponential machine intelligence will transform the physical inputs and outputs of civilization.
About Me
- Born and raised in Juneau, Alaska. Circa 1917, Juneau’s Treadwell mine was the largest hard-rock gold mine in the world. Juneau’s Mendenhall glacier has receded ~3,500 feet over my lifetime.
- Studied chemistry at Reed College to understand how the world works. In 2018, pivoted into simulation and AI to change it.
- Ph.D. in Chemistry and NSF Fellow at UC Berkeley. Discovered new battery electrolytes with Kristin Persson at Lawrence Berkeley National Lab.
- Rotated through startups and government: simulation and deep learning architecture at Chemix, tech-to-market at ARPA-E, computational materials at Radical AI, and (briefly) AI agent research for Jordan Belfort.
- Co-founded Riven with Chip Breitenkamp. We are building the future of industrial chemistry.
A Manifesto
Over the past decades the western industrial base has been hollowed out by globalization and financialization. Key material inputs flow through compromised supply chains, endangering global stability and prosperity. This is especially true in critical minerals, where whole swathes of the periodic table languish in insecurity. As the US seeks to reindustrialize, it faces a precipitous knowledge gap and disheartening timelines to new capabilities.
Simultaneously, we are amidst the greatest technological transition of all time. Digital industries are being transformed overnight by exponential machine intelligence. In the physical world, however, that transformation will not happen by itself. Engineering realities impose thermodynamic constraints on the transmutation of matter. Chemical processes are unpredictable and new capacity cannot simply be invented from first principles; there must be data, carefully curated. Thus, exponential acceleration in the chemical industries requires systematization: rapid experimental iteration across length scales, integration with advanced computational modeling, and the modularization of deployed capacity.
Ultimately, every abundant vision of the future requires abundant material inputs. The energy demands lithium and cobalt in batteries and neodymium in generators and motors. High-field fusion magnets rely on superconducting tapes made from yttrium, barium, and copper. Space exploration and autonomous machines extend the demand to specialized sensing materials: mercury cadmium telluride for spaceborne infrared detectors, indium gallium arsenide for lidar. Producing these materials at the required purity and scale, without incurring massive environmental disruption, demands rapid innovation in industrial chemistry. In this sense, our ability to build the future is bottlenecked by our ability to process raw materials.
Personal
I am broadly interested in the natural sciences; I formerly studied meteorites and discovered a new species of orchid (manuscript in preparation). I hold a black belt in Tang Soo Do. I live in a Zen center in New York City.