AI for Warfighters: What Silicon Valley Gets Wrong | Tyler Saltsman

Most people think AI is ready for modern warfare.
It’s not.
In this episode, I sit down with Tyler Saltsman, Founder and CEO of EdgeRunner AI, to break down what artificial intelligence can actually do on the battlefield—and where it falls short.
Tyler is building domain-specific AI models designed to operate directly on-device, enabling warfighters to make better decisions in real time without relying on internet connectivity.
We discuss:
- Why many commercial AI models refuse real-world military tasks
- The problem with bias, guardrails, and lack of transparency in current AI systems
- Why AI must operate offline in contested environments
- The gap between Silicon Valley and the warfighter
- The biggest bottlenecks inside DoD acquisition and procurement (ATO, OTA)
- How AI is changing tactical decision-making at the unit level
- The risks of relying on a small number of dominant AI labs
We also get into the realities of building a defense tech company today, and what it takes to move from prototype to fielded capability.
This is a grounded, operator-level conversation about AI in warfare—focused on execution, not hype.
---
About the Guest:
Tyler Saltsman is the Founder and CEO of EdgeRunner AI, a defense technology company building domain-specific AI models for military applications. He previously worked at AWS, where he was involved in large-scale AI model training and infrastructure development.
---
Subscribe to At the Water’s Edge for conversations with operators, policymakers, and builders shaping the future of national security.