Gill Verdon, a former quantum computing engineer at Google, is on a mission to advance large language models (LLMs) using an innovative approach: thermodynamic computing. His startup, Extropic AI, has raised $14.1 million to pursue this exciting technology.
Verdon’s journey began in the summer of 2022 when he decided to pivot from his career in quantum computing and quantum machine learning.
“I left my career in quantum computing, quantum machine learning, took a couple months to gather my thoughts, and I went for it,” he explained in a recent interview with Peter Diamandis.
The core of Extropic’s innovation lies in its unique approach to hardware design. While initial prototypes require cryogenic temperatures, the ultimate goal is to create room-temperature silicon-based chips. Verdon elaborated: “When you prototype things, you start with the biggest, most macroscopic prototype you can because it’s simpler, it’s easier to get going, and you can probe it and understand it better.”
One of the key challenges Extropic faces is embedding complex AI algorithms into the physical hardware. Drawing from his quantum computing background.
“My whole career was figuring that out — how to embed quantum physics into a quantum mechanical computer, how to embed AI algorithms into Quantom mechanical physics,” said Verdon. This expertise is now being applied to thermodynamic computing.
Verdon’s vision extends beyond just creating a proprietary technology. He aims to foster a collaborative ecosystem around thermodynamic computing.
“Frankly, part of my goal is to partially open source the concept of a theramic computer so a broader Community can join us on this Quest,” he stated.
As Extropic AI continues to develop its technology, Verdon’s urgency is palpable. His commitment to advancing AI through novel computing paradigms reflects a broader movement towards more efficient and powerful artificial intelligence systems. With thermodynamic computing, Extropic AI may be on the cusp of a significant breakthrough in the field of AI hardware.
