Barron Stone

Airplane faults don’t wait to be reproduced. AI can catch them before they count. I spent time with maintenance teams on GA aircraft this past week, and one thing struck me. Fault detection still relies far too much on trial and error. In one case, a minor malfunction was reported but couldn’t be reproduced during inspection, so the aircraft was cleared. When I flew it, the issue resurfaced, and straight back to maintenance it went. It made me ask if we can l do better? I believe that in the AI era, the answer may be YES. This article from the World Aviation Festival highlights why airlines and airports are still hesitant to adopt AI for turnaround management. It also highlights why that mindset needs to evolve: 🔗 https://lnkd.in/dHvSTwH2 AI can deliver real-time visibility, predictive insights, and earlier anomaly detection. These are capabilities human teams (no matter how skilled) cannot provide consistently. Predictive maintenance is especially compelling. By analyzing live operational data with AI models, we can identify emerging issues before they become operational disruptions. Instead of intermittent checks or waiting for a fault to reappear, AI can forecast component health and guide maintenance teams toward proactive, safer decisions. Thus, reducing downtime and strengthening reliability. The aviation industry is under real pressure. There are staffing constraints, growing operational complexity, and rising expectations for punctuality and safety. I do not think AI is a silver bullet. However, it can reduce avoidable disruptions and help teams work smarter. AI has the potential to transform how we manage turnarounds and maintenance. There is no better time for the industry to lean in. P.S. In the photo, I am with a beloved PA-44 Seminole. I simple love this plane: #Aviation #ArtificialIntelligence #PredictiveMaintenance #AviationSafety #AviationInnovation #DigitalTransformation #TurnaroundManagement #FutureOfFlight

43

2 Comments

Heading home from #NACON Droneresponders energized by the people and the progress our industry is making. At the same time, it feels like we’re at an inflection point. Public safety UAS programs are no longer experimental, they’re mission-critical. That means it’s time for the next step in professionalization: • Recognized Drone Program Manager certifications • Fleet Management certifications • Standards that move beyond “buckets” and toward validated competency • Structured roundtables where agencies directly shape manufacturer roadmaps Conferences shouldn’t just showcase innovation, they should help define the professional standards of our industry. I’d love to see us build that next layer together. #drones #publicsafety #conferences #dobetter

71

5 Comments

NASA’s latest open Software Catalog, released Wednesday, offers more than 1,200 downloadable codes developed by agency engineers that could enable faster solutions to energize the space economy and stimulate American ingenuity.  https://software.nasa.gov/

19

It takes time to see outcomes from investments in the industrial base. That’s why the best time to get started is yesterday, and the second-best time is right now. This is one conclusion to be drawn from the Navy’s remarks to Congress last week. Congress expressed concern with the Navy’s progress in meeting 2+1 submarine goals, despite $9 billion in appropriations since 2018. In response, the Navy pointed out significant production improvements in the targeted areas, but acknowledged it was not seeing the benefits manifest yet in faster ship rollout. This is unsurprising. Industrial base supply chains are complex and interdependent. It takes time to revamp production lines and see tangible benefits across the supply chain. It is not enough to simply purchase new technologies or train new workers. What’s needed is re-envisioning the end-to-end production workflows around new innovations so that overall throughput ratchets up. This takes time, so it’s critical to start as soon as possible if you want to see improvements on the tail end faster. https://lnkd.in/edTZhDVY

33

2 Comments

I’m not in the market for a move right now, but I’ll bump the Head of Autonomy role at True Anomaly to my network. If you have experience across the range of spacecraft autonomy domains - mission sequencing, modes & states, system fault protection, GNC, state estimation, decision-making under uncertainty, game theory, ML/AI, etc. - this role is for you. It kills me not to be able to get involved with this incredible opportunity because I love every aspect of spacecraft autonomy, but for now, someone who is far more experienced will have to lead the charge.

16

Hypersonics and responsive space aren’t constrained by ideas or modeling. They’re constrained by access to repeatable flight testing. Reusable, licensed flight infrastructure enables the cadence required to move from concept to capability. That’s where real learning happens, and that’s the layer we’re focused on building.

7

#CognitiveEW in San Antonio 23-25 June: https://lnkd.in/gabJYMd4 Other options for longer classes, if North America is too far: ► 12-14 May UK 3 day class https://lnkd.in/gFhM-jxT ► 21-22 July Singapore 2-day class https://lnkd.in/gNEgEGRv 𝗙𝗼𝘂𝗻𝗱𝗮𝘁𝗶𝗼𝗻𝘀 𝗼𝗳 𝗖𝗼𝗴𝗻𝗶𝘁𝗶𝘃𝗲 𝗘𝗪: Motivation of why we need AI to solve modern EW challenges. Cognitive EW definition and explanation of where to leverage AI. 𝗦𝗽𝗲𝗲𝗱 𝗼𝗳 𝗔𝗜 𝗶𝗻 𝗘𝗪: The global landscape of AI research, development, and deployment. 𝗠𝗟 𝗽𝗿𝗶𝗺𝗲𝗿 𝗮𝗻𝗱 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗳𝗿𝗼𝗺 𝗼𝗻𝗲 𝗲𝘅𝗮𝗺𝗽𝗹𝗲: Covers basic ML concepts, algorithmic trade-offs, and the benefits of hybrid learning approaches. 𝗦𝘆𝘀𝘁𝗲𝗺 𝗗𝗲𝘀𝗶𝗴𝗻: Objective functions; situation assessment for electronic support; decision making for electronic protect, attack, and battle management; real-time operations (including battle damage assessment); and software architecture. 𝗗𝗮𝘁𝗮 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁: Curation, data modelling, and management practice. Test & Evaluation of Cognitive Systems: Discusses trust as a function of risk, and differences between evaluating AI-enabled systems versus cognitive systems. 𝗦𝗮𝗺𝗽𝗹𝗲 𝗣𝗿𝗼𝗷𝗲𝗰𝘁𝘀: Tools and small-scale projects to get started. 𝗦𝘆𝘀𝘁𝗲𝗺 𝗪𝗮𝗹𝗸𝘁𝗵𝗿𝗼𝘂𝗴𝗵: Detailed discussion of a fielded system, including its architecture, algorithms, performance, and testing approaches. In the 3-day courses, 𝗣𝘆𝘁𝗵𝗼𝗻 𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴, with (1) Clustering, with feature selection and regularization, (2) Classification for Specific Emitter Identification, with augmentation, and (3) Scenario control for a decision maker that can do In-mission learning

22

1 Comment

Thrilled to shout out my friends at Absentia Technologies! Their Adaptive AI Vision systems cut through smoke, darkness, and weather to deliver verifiable insights for defense and commercial security. All via software that upgrades existing drones and surveillance systems to turn degraded footage into actionable intel. Their MYSTIC platform and AI Agents achieve over 30% clarity improvements without hallucinating details, making them ideal for critical infrastructure protection and autonomous ISR operations. With an extraordinary team of veterans and AI experts, they're poised to transform the defense and private security industries. Check them out at absentiatech.com and follow them on LinkedIn. Excited to watch them grow! #DefenseTech #SecurityTech #AIVision #ComputerVision #ISR #NationalSecurity #DualUse #EdgeAI #DeepTech #TechStartups

17

2 Comments

Laser comms, laser power, laser weapons - all come out of the same space infrastructure. The emerging space ecosystem will increasingly blur the lines between civilian and military applications. https://lnkd.in/gh49sH8B

2

The weaponization of drones is one of the hottest topics being discussed. Absent from these conversations has been the weaponization of data which will be used to deploy these drones. Consider how much data is generated and made publicly available every hour of the day. The real value behind autonomous systems is their ability to leverage data from which patterns can be derived, and thus used as a predictive analytics engine for these platforms. The recent White House EO on unleashing drone technology combined with the advancements of AI models will increase the risk of autonomous drone technology being used for nefarious activities. However, just as with every other technological advancement in the past, the solution is not to curb their use in society. Autonomous delivery, extended wireless communication networks, surveying operations, force protection and more, will become normal operations for drones. Organizations need to be thinking ahead to understand the risks and investment opportunities this presents in the very near future. #drones #autonomous #cybersecurity #defense #logistics

22

4 Comments

Best Defense Tech of 2025 — what moved the needle There’s a lot of genuinely futuristic tech on this list. What stands out is how much of it made it past slides and controlled demos. In 2025, contracts followed systems that: - Paired autonomy and AI with real operators - Delivered usable capability before being “done” - Solved problems people actually feel today (counter-swarm, edge targeting, undersea infrastructure) Excited to see where this momentum goes in 2026!

6

I’ve often used the SysML v2 Viewer integrated in Sensmetry's Syside or Planetary Utilities's Starforge in SysML v2 training sessions. I’m looking forward to talking to Brendan about the tool.

24

Proud of the team on this one. United States Space Force is not a customer you win by checking boxes. Space domain technology operates at the intersection of the most complex engineering challenges in defense, contested environments, extreme operational constraints, and mission stakes that don't allow for mediocrity. This Task Order represents something bigger than a contract win for Metronome. It's a signal that the technical foundation we've been building, the engineering depth, the mission focus, the willingness to take on hard problems is being recognized in domains where the bar is genuinely high. Getting a seat at that table doesn't happen by accident. It happens because a team of people made hard technical bets, executed with discipline, and built credibility one capability at a time. This is the next step in a longer arc. More to come.

14

1 Comment

My career on paper is weird. Thermal analysis. Propulsion system integration. Open software standards. Now business development. None of it was planned. But every step built toward my passion: finding tech insertion points and connecting capability across the Joint Force. That work just got a whole lot easier. AMS GRA and A-GRA went Distro A this week, joining UCI and OMS. WGRA and T-GRA are next. These standards let defense systems be built like modern software. Modular, swappable, vendor-agnostic. The CCA program has been proving it on real airframes all year. Shield AI’s Hivemind on Anduril’s Fury and Northrop’s Talon IQ. Collins’ Sidekick on General Atomics’ Dark Merlin and Anduril airframes. Even mid-flight AI hot-swaps on the Fury. Same standard underneath all of it. For years, “but how do we integrate it?” was where good ideas stalled. Now it’s the starting line. Bring the ones you’ve been sitting on. https://lnkd.in/eMhsRUiH

8

If you’re building in defense or space and feel like you’re pushing hard but not breaking through your issue is not speed. 𝐈𝐭 𝐢𝐬 𝐚𝐥𝐢𝐠𝐧𝐦𝐞𝐧𝐭. I developed the Adoption Alignment Map after seeing the same pattern over and over. A way to simplify a complex ecosystem - sure it might be an oversimplification but its meant to be a guide not a SoP. I see plenty of technically strong teams stall; not because the tech doesn’t work, but because maturity and customer preference aren’t moving together. Two forces determine whether you transition or tread water: 1. 𝐓𝐞𝐜𝐡𝐧𝐨𝐥𝐨𝐠𝐲 𝐑𝐞𝐚𝐝𝐢𝐧𝐞𝐬𝐬: Is your capability prototype, demonstrated, or operational? 2. 𝐏𝐫𝐨𝐬𝐩𝐞𝐜𝐭 𝐏𝐫𝐞𝐟𝐞𝐫𝐞𝐧𝐜𝐞: Is the accountable problem owner unaware, interested, or actively preferring your solution? When those aren’t aligned, you get friction. It looks like: • “Interesting” conversations that don’t convert • Pilots that don’t scale • Budget that doesn’t materialize • Decision fatigue The map isn’t about judging where you are. 𝐈𝐭’𝐬 𝐚𝐛𝐨𝐮𝐭 𝐥𝐨𝐜𝐚𝐭𝐢𝐧𝐠 𝐲𝐨𝐮𝐫𝐬𝐞𝐥𝐟 𝐡𝐨𝐧𝐞𝐬𝐭𝐥𝐲. Because early on, you don’t need more acceleration. 𝐘𝐨𝐮 𝐧𝐞𝐞𝐝 𝐨𝐫𝐢𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧. Over the next few weeks, I’ll walk through each quadrant: What it looks like. What traps founders fall into. And what actually moves you toward adoption. If you’re a founder with space technology or maybe even deep tech and trying to get it adopted by a US government customer — this framework may help. ℱ𝑜𝓁𝓁𝑜𝓌 𝒶𝓁𝑜𝓃𝑔.

7

Profound Change Requires Building a Broad Coalition—nTop Understands the Challenge The 3rd annual nTop Computational Design Summit earlier this week in Los Angeles was again an impactful experience. This is my third year in a row attending, and the events have been 3 for 3. The event can be described as inclusive yet intimate, and the new location this year allowed a better flow. There was enough attendance to provide impactful exchange of ideas without diluting thought leadership in those engagements, great presentations, and powerful networking. Innovation thrives in collaboration. nTop recognizes their keystone position in the computational design ecosystem and their importance to the future of AI-driven design. They aren’t losing focus from what they do well, nor becoming overambitious in their role in the flurry of AI-workflow acceleration concepts out there. Several partner companies working those angles exhibited in a handful of booths, selected for the value they are bringing to nTop enabled workflows as "plugin" tools. nTop’s strategic partnering with the industry is helping bring it along. The message is clear, AI-driven design demands robust geometry, and implicit geometry under nTop modeling is as reliable as it gets. Bradley Rothenberg has a clear vision for nTop's potential for Aerospace design and it has been a pleasure watching nTop under his leadership not just adapt to the industry needs, but do so with passion and dedication. The industry also clearly recognizes the direction. Every established aerospace prime had their most experienced Multi-Disciplinary Design and Analysis Optimization subject matter experts in attendance. I spoke to some of them about their current progress with implicit-driven vehicle design stacks, and while adoption is still growing, presence alone shows the recognition and growing interest. Several newer aerospace companies and emerging product startups across the spectrum of vehicle types from air vehicles to missiles were also well represented too, and I had great conversations with multiple CTOs about their direction and experience with nTop-backed computational design workflows. Their anecdotes backed up themes discussed throughout the day including agility, requirements impact understanding, and broad design solution space exploration in concept development. You can read more on one of their stories here: https://lnkd.in/gRE4aXcp I’m looking forward to doing more with implicit-centered design optimization workflows in the next year and hope I can share more at next year’s summit. Hope to see everyone there!

27

THE JOINT MEANING LAYER Every joint operation has comms plans, ISR schemas, data fusion frameworks, and decision cycles. But none of those elements matter if the interpretive substrate between them fractures under pressure. In every AAR I’ve studied - from joint exercises to AI-enabled targeting environments - the real failure didn’t occur in sensors or systems. It occurred in the interpretation layer between units, functions, and machines: • Divergent frames • Misaligned inference models • Drift in shared mental maps • Human–AI symmetry breaks • Meaning collapse under tempo We keep optimizing systems for speed, precision, and scale. But we don’t optimize the meaning layer that makes them coherent. Until we stabilize interpretation across echelons, jointness is an illusion and integration is theater. The next decade of doctrine won’t be written around tools. It will be written around cognitive integrity - the true center of gravity in AI-enabled operations.

2