The Stack Overflow Podcast Robots in the skies (and they use Transformer models)
Aug 22, 2025
Ryan welcomes Nathan Michael, CTO at Shield AI, who's at the forefront of AI for defense technologies. They dive into Hivemind's tech that enables drones to make autonomous decisions while keeping humans in the loop. Nathan discusses the delicate balance between human oversight and autonomous operations, highlighting the ethical implications of their innovations. The duo also explores the challenges of operating drones in complex environments, emphasizing the importance of communication and adaptability in high-pressure scenarios.
AI Snips
Chapters
Transcript
Episode notes
From Robotics PhD To Shield AI CTO
- Nathan Michael began his career drawn to robotics and got a PhD from the University of Pennsylvania.
- He worked at Carnegie Mellon Robotics Institute before joining Shield AI as CTO to scale autonomy work.
Hivemind Is Autonomy + Factory
- Hivemind is both on-edge autonomy and an 'autonomy factory' for designing and deploying capabilities.
- The stack enables perception, decision-making, and deployment across multiple platform types.
Use AI Only Where You Can Assure It
- Shield AI frames work as AI-enabled autonomy, using AI only where appropriate and provable.
- They pick methods that give probabilistic bounds or guarantees before deploying to the edge.