Summary

Phillip’s central thesis is that aviation’s staffing, infrastructure, and technology failures are interconnected software problems. Intuitive tools can shorten training, reduce “10,000 clicks,” improve operator productivity, and potentially support higher pay—turning modernization into both a capacity and talent lever.
The procurement model is producing obsolete systems before they launch. Phillip Buckendorf describes tens of millions spent on thousand-page requirements, followed by hundreds of millions and 10 years of custom development, then 20 years of maintenance. His alternative is commercially proven software separated from compute, because “software is never complete.”
Logistics is a flagship dual-use case, though not every defense capability is. Government and industry rely on overlapping commercial transport, shipping, rail, trucking, and port capacity. ASI says airline-proven software helped it deploy with the Air Force “within months”; shared software and data can let public and private operators coordinate before a crisis. Leo’s NATO experience adds that allies increasingly need jointly planned “collective logistics” because no nation can do it alone.
Logistics is not overhead but a weapon system and competitive advantage. Leonard J. Kosinski’s test is blunt: advanced equipment that cannot be moved and sustained provides neither capability nor deterrence. Phillip sharpens it further: if it cannot arrive “where it’s needed, when it’s needed, it doesn’t exist.”
The next software layer moves beyond dashboards into prediction and forward simulation. “Anticipation in many ways is a new high ground”: operators need to see disruptions hours, days, or weeks ahead, while AI decision tools can support multimodal choices and reroute capacity before weather, sanctions, cyber threats, or conflict disrupt the network.
The urgency comes from vulnerabilities across ports, pipelines, water systems, undersea cables, and allied infrastructure. America’s two oceans make direct attack on the homeland difficult but complicate global power projection, while adversaries study supply-chain vulnerabilities and try to place free software in allied ports. The strategic objective is deterrence: “The best type of war you fight is one that you don’t have to fight at all.”

Deep dive

1. ASI began when transportation outsiders found ancient software behind critical operations

Phillip traces his departure from Germany to around 2011–2012: the country’s decision to exit nuclear energy and depend on cheap Russian gas, the top 5% of his friends and social network glorifying consulting, and a government that made starting companies and accessing venture capital difficult. Silicon Valley offered the opposite ethos—three years of bunk beds, Trader Joe’s frozen food, and people “obsessed with building and technology.”
He and K.D. Lucas came from autonomous driving, not piloting, defense, or logistics. When that market felt overcrowded in 2017–2018, they spent six months visiting maritime and air operations centers, expecting science fiction but finding “the most ancient software possible.” That discovery became ASI’s mission: enable critical operations and optimize the country’s most valuable assets and infrastructure.
Leo arrived after more than 30 years in the military, eventually retiring as a three-star general. His experience spanned industrial engineering and neural-network work, mobility aircraft, logistics for U.S. Africa Command during the pandemic, a command assignment in Japan, and Joint Staff logistics during support for Ukraine and Israel. His recurring Department of Defense problem was accessing and understanding data well enough to optimize what needed to happen.

2. Aviation’s labor and infrastructure problems converge in software

Phillip divides aviation’s predicament into staffing shortages, faltering legacy software, and outdated physical infrastructure—but rejects treating them as silos. Better software reduces workload, accelerates training, assists decisions, and raises each operator’s productivity; that improved productivity could support better compensation and make the field more attractive.
The generational mismatch compounds the shortage. The generation retiring or nearing retirement grew up with IBM green screens, while incoming 25-year-olds grew up with iPads, Snapchat, and Google Maps. Phillip argues that modern software is essential for bringing the tools up to the standards this new generation expects.
Leila framed the tension with aviation’s strong safety statistics and legacy technology, citing the recent Reagan National commercial-airline/military-helicopter collision. Phillip’s answer focused on a software philosophy that “might be a bit broken”: safety performance does not eliminate the problems created by tightly coupled, difficult-to-update systems.

3. Hardware-style procurement makes software obsolete before deployment

The first architectural failure is coupling software to local compute. Updating facilities across the country can require visiting them individually, with every patch constrained by whether the compute can handle it. Modernization must separate the two because today’s replacement “is not going to be the last update.”
The second failure is building software as though it were hardware. Phillip’s archetype begins with a thousand-page specification and tens of millions spent before anything works; custom development then consumes hundreds of millions and roughly 10 years. The finished product is already antiquated, yet may be maintained for another 20 years.
That structure repels top engineers, who want to build rapidly and close to users—not receive “10,000 requirements” and code against them. Leo sees the same pattern across the Army, Marine Corps, Air Force, and subordinate systems: capable people built tools decades ago that were never designed to connect, leaving the institution to improvise integrations. Phillip adds that an acquisition framework reinforcing these philosophies compounds the problem.
Phillip says momentum must hold: he credits President Trump and Secretary Duffy with setting direction, then calls for Congress to fund FAA modernization and establish spending rules that prevent another decade-long custom program. Where commercially deployed software already works, buying it largely as-is—with limited modification—is faster and, in his framing, safer because it has already been proven in continuous operations.

4. Logistics turns commercial capacity into military power

ASI’s air-traffic-management capabilities already operate with major airlines, Phillip says, while its commercial foundation allowed deployment with the Air Force “within months” and use in live operations. The software had already been deployed in a “24/7, 365 days a year” environment rather than beginning as a government specification.
Leo defines contested logistics through an ordinary delivery: consumers notice price and arrival time, rarely the supply chain underneath unless weather or another disruption causes a delay. Moving an aircraft carrier or soldiers also requires the unseen “tail”—food, munitions, replenishment, and everything else needed to get them there and back. Adversaries deliberately target those vulnerabilities.
The pandemic briefly exposed that dependency through masks and toilet paper, but Leo says human nature pulls institutions back toward comfort. Decades of just-in-time optimization treated logistics as a cost; under disruption, “just in time isn’t in time at all.” Like electricity or water, logistics is easy to ignore until it stops working. Resilience can instead become commercial advantage, military capability, and deterrence.
Phillip distinguishes purpose-built systems such as missiles and aircraft carriers from logistics, which he calls the “truest form of dual use.” Defense capacity already resides in commercial trucking, rail, aviation, shipping, and ports. Similar software and data structures let both sectors coordinate before emergencies—especially important when, he says, adversaries such as China try to place free software in allied ports for strategic reasons.

5. Collective logistics and prediction machines unlock existing capacity

Leo’s NATO experience exposed an institutional gap: “There’s always been collective defense, but not necessarily collective logistics.” Nations historically carried their own burden, but NATO has moved toward a common understanding and joint planning because any one nation cannot do it alone and must coordinate allied infrastructure and capacity.
Phillip argues that recent modernization often amounted to new dashboards on larger operation-center televisions. Near-real-time awareness lets operators react as problems occur; a more uncertain world requires software that shows what is about to happen and how operators need to adjust. “Anticipation in many ways is a new high ground.”
His technology arc runs from isolated 1970s–1980s workstations, through networked computers and internet-connected sensors, to roughly 15 years of data fusion and common operating pictures. The emerging layer is “prediction machines”: forward-simulating asset, supply-chain, and operating conditions across hours, days, and weeks.
Leo applies that capability to military expertise. Becoming proficient in one service’s logistics takes several years, if not a decade; becoming a joint logistics expert across the Army, Air Force, Navy, and Marines takes much longer. AI decision tools could compare multimodal options such as ship versus air and accelerate decisions without pretending training no longer matters—or making success depend on one irreplaceable expert.

6. Resilient logistics is the foundation of deterrence

Ten years out, Phillip expects better software to extract more capacity from existing networks and harden them against disruption. His forecast remains hedged but stark: the probability is “very high” that coming decades will be more dynamic and uncertain than the last two, with geopolitical tensions, sanctions, and volatile weather affecting mission-critical infrastructure.
The desired system anticipates disruption, reroutes flows quickly, and prevents uncertainty from affecting warfighters or civilian infrastructure. Leo’s deterrence logic is equally direct: ships, aircraft, and advanced weapons matter only if they can be moved and sustained; otherwise, they do not provide effective capability or deterrence.
Geography cuts both ways. Two massive oceans make direct attack on the U.S. homeland difficult, Phillip says, but global power projection requires crossing vast distances. Adversaries therefore target key logistics infrastructure in the homeland and among allies, while software helps determine whether American equipment can arrive on time.
Leo cites The Hundred-Year Marathon by Michael Pillsbury while arguing that adversaries have studied supply-chain vulnerabilities for a long time. He also points to hacking of water-supply systems in Texas. Phillip adds the Colonial Pipeline disruption, the 2021 Suez Canal blockage—where a ship remained stuck for six days and billions of dollars of trade were affected—and attacks on undersea cables as examples of network fragility. Every key node is a vulnerability: detect threats, understand their profiles, and shift capacity toward less-affected nodes.

Episode 139 - April 25, 2025 Plural Partner, Taavet Hinrikus: Why Founders Will Realise Multi-Stage Funds Damage Seed Rounds