Defense Tech

Advanced violence is democratizing.  AI, in conjunction with dramatic improvements in robotics, energy production, and sensors, will increasingly enable ever-smaller groups of people to use targeted violence more effectively, and from a distance. Over time, this shift will dramatically impact all varieties of force projection: state-on-state war, various forms of low-intensity conflict, and how states enforce internal order. 

Perhaps understandably, however, national security discourse about the AI revolution has generally focused on more earth-shattering scenarios: superintelligence, state-to-state conflict, and the prospect of unleashing new biological weapons. These are all critical questions that deserve extensive scrutiny. But super-empowering small groups of people will shift security dynamics in crucial, if less dramatic, ways as well. Non-state actors will use AI-backed tools to conduct relatively simple attacks using increasingly autonomous weapons. In this scenario, it will be the ability of AI-empowered weapons to deliver destruction discriminately, rather than at a catastrophic scale, that will be critical. 

In the age of AI, the scarcest resource in headquarters is no longer time. It is, rather, the willingness to say no.

Artificial intelligence is moving rapidly into military planning staffs because it compresses routine cognitive labor. AI excels at absorbing guidance, reorganizing complex material, and producing clear strategic language at speed. This feels like a qualitative advance, creating the impression that planning itself has become easier. But this impression misleads. The risk of AI-enabled planning is that it will produce plausible constructs that obscure where judgment is required, creating the illusion that analytic completeness can substitute for prioritization.

AI is seen as “raising the floor” by making it easier to produce adequate products. That is true. Yet AI also “collapses the median” by increasing the relative cost of real insight. As AI-enabled planning begin to inform real-world operations, the temptation is to treat complete answers as sufficient, without interrogating whether they represent the right answers to the hard questions of what to resource, what to defer, and what risk to accept.

Agentic AI is quickly moving from demo to deployment inside the Department of Defense. But what does it actually mean to give AI “agency” — and what does it take to make those systems work on real military networks?

In this episode, Ryan sits down with Ben Van Roo, co-founder and CEO of Legion Intelligence, Jags Kandasamy, co-founder and CEO of Latent AI, and Aaron Brown, co-founder and CEO of Lumbra AI, to discuss why the real challenge is not just building smart models but getting AI agents to run on military networks and inside operational workflows. They cover deploying agents in denied environments, compressing models for the edge, orchestrating them across stovepiped systems, and the Pentagon’s struggle to scale and buy these tools fast enough to matter.