Callum Sharrock
Callum Sharrock is a Member of Technical Staff at Andon Labs, where he focuses on safety evaluations for frontier AI systems. That means deploying AI in real-world environments, figuring out every way it can go wrong, and reporting those findings publicly so organizations have an honest picture of where the technology actually stands.
He came to AI safety through robotics, with hands-on experience spanning welding defect detection in Berlin, programming robots to repaint digital art in real acrylic on canvas, a year at Tesla working on the Robotaxi cleaning robot and camera systems for Autopilot and Optimus, and reinforcement learning research for compute-constrained robots at the Vector Institute.
His view is that robotics is now a software problem, not a hardware problem. And that most people are underestimating the trajectory of AI because they're looking at a snapshot instead of the trend line.