This episode explores how autonomy has transitioned from a 'moonshot' concept to a tangible market force, detailing Applied Intuition's dual-use strategy leveraging commercial advancements to drive innovation across commercial and defense sectors.
Introducing Applied Intuition: Bridging Commercial and Defense Autonomy
- Applied Intuition positions itself as a vehicle intelligence company, developing software and AI—including tools, operating systems, and autonomy solutions—that are integrated into vehicles. Co-founder and CEO Casser Unus explains the company, founded 8-9 years ago, is based in Mountain View, CA, with a significant DC presence established 5-6 years ago, now employing around 1,000 people, primarily engineers.
- Jason Brown leads the government business from the DC office, which houses about 110 dedicated staff, though the entire company contributes to government projects. He emphasizes their role as an "all domain autonomy company," working with every US military service and expanding internationally to allies, with projects spanning air, maritime, and land domains.
Casser Unus' Journey: From Automotive Roots to Leading Vehicle Intelligence
- Casser Unus shares his deep roots in the automotive industry, starting with an undergrad degree from the General Motors Institute and early career roles at General Motors and Bosch. He notes his involvement with early ADAS (Advanced Driver-Assistance Systems – technologies that assist drivers, like cruise control or lane-keeping) at Bosch around 2004-2005.
- His interest reignited during his time at Google (post-acquisition of his startup in 2011) when the self-driving project began. He later connected with co-founder Peter Ludwig (founding engineer on Android Automotive), bonding over the intersection of software and vehicles.
- They founded Applied Intuition recognizing that the dynamics driving autonomy in automotive—like software integration and intelligence—were mirrored in defense, construction, mining, and trucking, leading to the company's natural expansion across land, air, and sea applications.
The State of Autonomy: From Sci-Fi Speculation to Production Reality
- Casser observes a significant shift in the perception and reality of autonomy over the past decade. Initially viewed as speculative, almost sci-fi ("like crypto," he jokes, with cycles of excitement and letdown), autonomy is now entering a phase of practical application and scaling.
- He points to the tangible presence of services like Waymo as proof of concept. Casser asserts, "the next decade is very much the productionization proliferation of autonomy... and it is also the monetization of this technology." Applied Intuition focuses on this "applied side of AI," bringing tested research to mass markets.
Two Paths Converging: Waymo vs. Tesla and the Future of Autonomy
- Casser outlines two prevailing, though simplified, approaches to autonomy today:
- Waymo: Fully driverless (no human in the driver's seat), sensor-heavy (expensive hardware and compute), proving technical feasibility but still scaling the business model. Achieves high reliability (tens of thousands of miles between disengagements).
- Tesla: Driver-assist (human driver required), sensor-light (primarily cameras), with a proven business model (paid FSD - Full Self-Driving, Tesla's advanced driver-assist feature suite), but requires driver oversight due to limitations.
- He predicts a technical convergence over the next decade. Waymo aims to reduce costs and generalize, potentially looking more like Tesla's system, while Tesla works to improve reliability and reduce driver dependency, moving closer to Waymo's capability. This convergence pattern is expected across other vehicle types like trucking and defense.
Beyond Passenger Cars: Autonomy's Expansion Across Industries
- Casser emphasizes that the fundamental technological shifts and business dynamics driving autonomy in passenger cars are equally relevant in commercial trucking, defense, construction, and mining.
- He notes that traditional automotive OEMs (Original Equipment Manufacturers) globally are adopting models more akin to Tesla's approach – focusing on intelligent in-cabin experiences, driving assistance, and connected ecosystems (referencing Xiaomi's smart home integration demos in China). The Waymo model is seen more as a direct replacement for ride-sharing services like Uber.
The Autonomy Stack: Abstraction, Transformers, and Cross-Domain Application
- While no single "standard" autonomy stack exists, Casser highlights a key trend: increasing abstraction away from specific hardware. This allows for greater software portability across different vehicle types and sensors.
- He reveals that Applied Intuition's autonomy system for a defense vehicle is roughly "90% similar" to their commercial passenger car or trucking solutions, a level of commonality not possible just a few years ago.
- This progress is partly driven by architectural shifts, particularly the use of Transformers (a type of neural network architecture highly effective at processing sequential data, famously used in Large Language Models) in vision and autonomy systems. Similar to LLMs, Transformers enable better scaling with large datasets in autonomy, facilitating what's often termed "end-to-end" learning approaches.
Defense Autonomy Challenges: Long Lifecycles and Retrofitting Legacy Systems
- Jason Brown points out a critical challenge in defense: the extremely long operational lifecycles of military hardware (citing the B-52 bomber potentially reaching 100 years of service). This necessitates strategies for upgrading existing ("legacy") systems.
- He explains that software is key to making these older platforms "more lethal, more mobile, more survivable." Applied Intuition employs similar development approaches used for new commercial vehicles to enhance legacy defense systems.
- About 50% of Applied Intuition's defense work involves retrofitting or enhancing these legacy platforms. Jason candidly states, "we have to jailbreak a lot of these systems... show us how you can actually make this a better vehicle because it's it's really difficult to do it any other way."
Applied Intuition's Defense Genesis: From EBC Meeting to DoD Engagement
- Casser recounts the company's entry into the defense sector around 2019, spurred by an A16Z Executive Briefing Center (EBC) meeting with defense prime contractor General Dynamics. Showcasing their software tooling (used for data ingestion, curation, model building, and testing) resonated deeply, revealing shared challenges between commercial auto and defense primes.
- Initial engagement involved navigating the complex defense ecosystem, including working with the Defense Innovation Unit (DIU) and leveraging programs like SIBRs (Small Business Innovation Research – government programs funding R&D in small businesses). Jason identifies the Army's Robotic Combat Vehicle program as their first major defense contract.
- Casser advises founders entering defense to "just have to start," finding pragmatic ways (like SIBRs or partnerships) to gain initial traction and build momentum.
Market Dynamics: Shaping the Defense Market with Commercial Benchmarks
- Jason notes that while the commercial autonomy market is now pulling technology into production (using Waymo as an example), the defense side still requires significant "market shaping."
- Casser uses the analogy of the mobile revolution: the DoD eventually needed a mobile strategy but couldn't realistically design and build smartphones competitive with Apple or Android. Similarly, defense needs to leverage commercial autonomy advancements.
- He argues that dual-use companies like Applied Intuition provide crucial value by grounding defense work with commercial benchmarks. "We always get like baseline to the truth. If our products are not good, we're not going to survive as a company," Casser states, highlighting the competitive pressure of the global commercial market.
The Power of Dual-Use: Commercial R&D Fueling Defense Innovation
- Jason emphasizes that the nature of autonomy is inherently dual-use, with commercial benchmarks currently leading development standards (ROI, iteration cycles, etc.).
- Crucially, a feedback loop exists: innovations developed for defense (e.g., off-road autonomy for the Army) directly benefit Applied Intuition's commercial applications in agriculture and mining. This "flywheel effect" is expected to accelerate as they expand into aviation and maritime domains.
- Casser adds that software development inherently compounds; features developed for one customer or domain benefit all customers across both commercial and defense, amplifying the impact of R&D investment.
Technical Nuances: Adapting Core Technology for Diverse Domains
- While the core technology stack is largely shared (estimated 90%+), adaptations are necessary for specific domains. Jason provides the example of "Maritime Sim," a simulation environment for naval applications.
- Leveraging their existing commercial simulation tools (built with an estimated $750M in engineering investment), Applied Intuition created the first functional version of Maritime Sim in just 5 weeks, incorporating unique domain physics like the curved horizon.
- Casser clarifies that while the vast majority of the tech is interoperable, optimization occurs for specific use cases, especially highly classified defense projects. The key takeaway, Jason stresses, is that the core R&D is funded by the commercial sector, not solely the taxpayer.
Organizing for Dual-Use: Structure, Talent, and Security Considerations
- Applied Intuition maintains a dedicated DC office with government expertise, but most engineers work across both sectors. Casser acknowledges the need for structured processes to handle classified work, especially given their diverse, international workforce (including Chinese and Russian nationals).
- However, he pushes back against the notion that defense work requires complete segregation, stating "there's a lot more overlap than we think." Jason adds that much development can occur in unclassified environments before being deployed securely, aligning with DIU's goal of bridging Silicon Valley and the DoD. Engineers in Mountain View might work on commercial OEM projects in the morning and Air Force projects in the afternoon.
Modernizing DoD Procurement: Progress, Hurdles, and the Need for Software Acumen
- Assessing DoD's tech procurement modernization, Casser offers a nuanced view: zooming out, there's clear improvement over the last five years. However, he argues the biggest lever isn't process bureaucracy, but rather having senior DoD leaders with genuine, hands-on software development experience who can cut through red tape and accurately assess technology.
- Jason provides a more critical perspective, noting that the top 98 venture-backed DoD contractors (excluding SpaceX and Palantir) account for only ~$10B/year, equivalent to one aircraft carrier—a tiny fraction (around 1%) of the DoD's $800B+ budget. Challenges include a hardware-centric focus and difficulty distinguishing true software product companies (which scale efficiently) from services companies (which scale linearly with headcount).
- While initiatives like the recent Software Pathways memo (guidance aimed at streamlining software acquisition) are positive steps ("a beam of light"), Jason cautions that enforcement is key, and the services themselves, not just DIU, must learn to procure software effectively.
Navigating the DoD Landscape: Advice for Startups
- Casser advises "radical pragmatism": use whatever channel (direct relationships, DIU, prime partnerships) gets you into the game fastest. However, he stresses the difficulty of building a sustainable government business without a dedicated DC presence staffed by people with deep ecosystem experience.
- Jason strongly recommends startups always build direct relationships with the end government customer (users, buyers) regardless of partnerships. While partnerships (with primes, other startups, etc.) are valuable, they require significant effort and shouldn't replace understanding the government's direct needs. Sometimes, a smaller dual-use company can open doors even established partners cannot.
Autonomy's "ChatGPT Moment": Consumer Awareness vs. Industry Certainty
- Asked if autonomy has had its "ChatGPT moment" (a point of broad public awareness and appreciation), Casser distinguishes:
- No: For the average consumer outside tech/defense hubs, autonomy isn't top-of-mind like ChatGPT. Access is limited (buying a car vs. visiting a webpage).
- Yes: Within the automotive and technology industries, there's "no debate" about autonomy's inevitability. Every global OEM is actively developing solutions.
Safety Criticality: The Defining Difference in Autonomy Development
- Casser underscores a fundamental distinction: unlike chatbots, vehicle autonomy is inherently safety-critical. He draws on his safety engineering background, stating driving is "the most dangerous thing that you do by far every single day."
- This reality necessitates rigorous development, testing, and validation, contributing to longer development cycles compared to purely digital AI applications. He sees autonomy as crucial for saving lives and reducing accident-related costs, predicting significant impact within 5-7 years.
The Global Proliferation of Autonomy: Beyond Consumer Vehicles
- The panelists agree that autonomy's impact extends far beyond personal cars into industrial applications (construction, mining) and defense. Every major vehicle OEM globally is investing in autonomy and related modern operating systems.
- Jason highlights the rapid democratization of drone and autonomy technology, citing examples from Ukraine and the Middle East (Yemen). He notes inexpensive, AI-enabled systems being used effectively by less sophisticated actors like the Houthis against advanced military assets.
Geopolitical Implications: Democratized Drones and the Rise of Collaborative Swarms
- Jason, drawing on his background as Director of Intelligence for Air Force Central Command, analyzes the current geopolitical landscape. While current drone attacks (e.g., by Houthis, or against Israel) often involve uncoordinated assets, the underlying technology is accessible.
- The critical next step, he argues, is collaborative autonomy – effectively networking autonomous systems (like drone swarms) to behave collectively. He states, "that is inevitable... the first time that that collaborative swarm is used anywhere... that is just going to completely upend... the entire defense industry globally."
Applied Intuition's Future Focus: Culture, Talent, and Technical Execution
- Looking ahead, Casser emphasizes maintaining Applied Intuition's deeply technical culture, ensuring engineers are productive, focused on the right problems, and out-innovating competitors. Recruiting top technical and business talent remains a key priority ("we're hiring").
- Jason highlights the company's blend of "curiosity-driven engineers" tackling hard technical problems and "mission-driven folks" focused on delivering capabilities to warfighters and allies. He finds the synergy between these groups, particularly when Silicon Valley engineers collaborate directly on defense projects, incredibly powerful and central to their future direction.
Reflective and Strategic Conclusion
Autonomy is now a tangible market reality, driven by Applied Intuition's powerful dual-use flywheel where commercial R&D directly fuels defense capabilities. Crypto AI investors and researchers must track this commercial-defense convergence and the strategic implications of emerging collaborative autonomous systems for future market and security landscapes.
