Anduril Observer

@mikenelson586 Good info, well written article. But... aren't we well past debating whether autonomy is part of the future of warfare? The more pressing questions now are about acquisition, doctrine, training, and production—how do we actually integrate these systems into the force at scale?

@reoindustries As long as it's optional, this doesn't seem like it's in conflict with the vision you've been marketing. There will be a few with knee jerk responses to this. I suggest you not waste time using logic to talk them down from an opinion they didn't use logic to come to.

What we’re about to tell you will be completely OPTIONAL and won’t affect the core truck. It will feel controversial at first, but it’s crucial to understanding REO long term and how we’re so much more than our promise to bring reliable and affordable cars back in American hands. Just hear us out: Long term, our intent is to revolutionize FSD, and vehicle modularity in general, by developing a program that gives every REO the OPTION to be self-driving post launch. How can we do this while keeping our freedom approach? Well, keep reading. Every REO ships as bare bones as possible. No screen slop, no spyware, no telematics. All freedom to the driver. That said, some people want more tech and lots of people want self-driving vehicles. So we’ll give them the OPTION. For fleet buyers especially, this makes a ton of sense. That’s why our roof is designed for modular attachments. Lighting, racks, camera systems, and anything else you can think of, including LiDAR/Radar units front and rear. Our intent is to develop a program where an aftermarket FSD package can be fitted directly to the roof mounts. That same program would include a full retrofit kit for steering and brake components, bolted on aftermarket, to turn any REO into an autonomous vehicle for both hardware and software, all through our marketplace. Again, this will be an OPTION only. The core truck will always be as simple and freedom oriented as federal standards allow.

It always feels like this technology has been “a couple years away”, but it’s finally starting to look practical. If lasers can reliably handle drones and cruise missiles, they don’t replace traditional interceptors—they preserve them for the threats that actually require a that particular solution. That could dramatically reduce both the time and cost of rebuilding the arsenal AND make those inventories last much longer in a conflict.

I think you're right and the key words are "civil service". Get the individuals the right kind of job for them (based on competence, not preference). I've heard people say that the military should provide this function. Let me say in a loud clear voice, the last thing our military needs is to deal with a huge influx of people who don't want to be there.

I’ve been arguing for over a decade that mandatory civil service would make America stronger. The typical response was that it would just be an obstacle between people and their goals…or that it’s uNConStITutioNaL and that we’d rather have violent tribalwarfare between the purple hair baristas and the incels. Now that widespread job loss driven by AI looms, the idea has become a lot more palatable. Go check out the argument @PirateWires. Come at me, bro.

@BrettKrieger12 It's going to be really interesting

Software Sold Separately: What the Air Force’s New Approach for CCAs Means Shield AI, Anduril and Collins Aerospace were selected to compete for the mission autonomy software that would allow CCAs to carry out operations without direct human control In 6 months, the Air Force will choose one or two vendors to continue for six more months, and in summer 2027, it will choose a single vendor @shieldaitech will be shaping Hivemind based on many lessons learned from a variety of Pentagon autonomy programs, including the Alpha Dog flight trials with an F-16

@DeepValue47 @PalmerLuckey I'm sure he could get a lot of people with plenty of funds at their disposal very excited by claiming otherwise, but @PalmerLuckey is a straight shooter.

Anduril founder Palmer Luckey @PalmerLuckey views facial recognition as too unreliable and spoofable for high-stakes, split-second lethal decisions in kinetic targeting (e.g., autonomous weapons systems that decide who lives or dies), so Anduril has an internal policy against integrating it for that purpose, despite acknowledging its value in lower-stakes investigative uses like law enforcement “Like biological weapons I wouldn’t build. Facial recognition is actually another one. I, I’m not, I’m not saying facial recognition has no uses. However, we’ve had a kind of internal policy, me and Brian Schimpf, our CEO, that facial recognition is probably the wrong tool to use as part of the kind of final stage targeting of kinetic weapons. It’s, it’s, it’s just still a very fraught technology. There’s a lot of ways to spoof it. There’s a lot of ways that it can be wrong. There’s a lot of ways that people can affirmatively trick a system that is built on facial recognition. And so to me, it’s, it’s a great tool for, let’s say, law enforcement to perform investigative activity. It isn’t a good tool to use when you’re deciding who lives and who dies in a split second decision. So we’ve, we’ve never integrated facial recognition into any of our systems. Because we don’t think the gains are even close to the, the, the negatives.” @demarest_colin / Axios

As if my dog doesn't lose his damn mind enough now when we get a package delivered.

Those are certainly concerns. The software isn’t going to be pushed straight from a developer’s laptop onto operational aircraft like a phone update. It still has to go through extensive testing and certification before it’s fielded. On the contracting side, the goal isn’t to rotate vendors just for the sake of it. It’s to avoid being locked into one company’s software for the next 30 years if someone else develops a better solution. More importantly, it avoids one of the biggest problems in defense acquisition: having to buy an entirely new platform just to gain a new capability.

Or planes will just drop out of the sky every time there's a bad software update. The real reason to do this is to make a crap ton of money by rotating contractors and contract requirements every few years. They get to update fees and charges every few years as the just bounce the contract back and forth among each other.

This feels like one of the biggest acquisition changes we’ve ever seen. Instead of buying an airplane and living with its software for decades, the Air Force wants continuous competition on the autonomy. Think about your smartphone. It keeps getting better through software updates without you needing to buy a new phone. Now imagine if you could also switch to a completely different operating system—one built by another company—without replacing the hardware. That’s what the Air Force is trying to achieve. If this works the way the Air Force hopes, they’ll never have to choose between keeping a proven aircraft and getting the latest autonomy. They’ll be able to have both. airandspaceforces.com/software-sold-…

@Johnny_Joey Make sure Marc picks up that tab!

Multiple generations of Marine EOD.

Who needs Hellfires when you've got Barracuda 100M's? The Barracuda-100M meant to occupy a similar employment profile as Hellfire, while extending many of its capabilities. Similarities: - Comparable size and price point - Compatible with existing launch infrastructure - Same general mission/target set Differences: - 120+ n. miles of range vs roughly 4-6 n. miles for Hellfire - A turbojet instead of a rocket motor - Lattice mission autonomy, collaborative behaviors, and software-defined upgrades - The ability to be launched from air, ground, or maritime platforms, not just helicopters

@AndurilObserver WHOA!!! Talk about a 1-2 punch! Speaking of the software, I started imagining the Fury carrying a pair of Hellfire missiles under each wing like the Reaper; same rails and everything! Now you have a high speed attack drone with the flip of a software switch.

@OriflammeTech No, that's not what I meant. But if that version of the analogy helps you, you're welcome to it.

@AndurilObserver You mean a personal computer, not a smartphone

Interesting and well written piece on the tradeoff between acquisition speed and independent testing. The sections on DOT&E and how AI-enabled systems create new testing challenges are a major concern and worth thinking about. It lost me on using Barracuda as the proof that oversight has already failed. We don’t know what classified testing the Army has already conducted, and public disclosures are almost certainly an incomplete picture. Example: The “70 wrong impacts” argument is speculative. It assumes: - Fixed failure rates - Every failure equals civilian impact - Every miss is due to guidance failure It's more complicated than that. A failed missile could: - Self-destruct - Miss harmlessly - Fail to launch - Abort - Lose propulsion - Impact an unoccupied area The real question is: How to we preserve independent testing without falling back into a procurement system that takes 10-20 years to field capability?

@EODstu Who knows?! I'm partially colorblind. 🤫🤐

@AndurilObserver Did you cut the red or blue wire? And why was it the red wire ?

I almost made it to my goal of 1,000 followers by the end of the weekend! As a thanks to all my new followers, here's picture of me prosecuting an IED in the Sangin District, Helmand Province, Afghan circa 2012. I love talking about the job. If you have questions, fire away.

@BellFlight Sexiest attack helo ever. Makes the Apache look like the fat friend who won't let her girlfriend talk to you at the bar.

Built for reliability, survivability, and mission success, the AH-1Z Viper stands ready to support warfighters in any environment. #AH1Z #Viper 📸 Lance Cpl. Bryan Giraldo 2nd Marine Aircraft Wing The appearance of U.S. Department of Defense (DoW) visual information does not imply or constitute DoW endorsement.

I still occasionally facilitate training with military EOD and LEO Bomb Squads. The thing that's made the biggest impact recently is using small UASs to do complete the initial recon of the item in question. It used to take us 30-45 minutes to get that accomplished, now it takes 5-10 minutes. They've even started putting some rudimentary tools on some these UASs to take a shot at rendering safe the item immediately. I can't to see where that goes as the systems get more and more capable. If we had had those options back during GWOT it would have saved a lot of lives and a lot of time.

@AndurilObserver In what aspects of the job do you think that unmanned systems available today would have helped most? What are your thoughts regarding doctrine and implementation efforts? Thanks a lot man!

@NotTodayTaliban We tried to convince the bad guys to just throw their IED components in our disposal area outside the FOB and we'd go out there and get rid of them occasionally, but they kept assembling them and burying them in random places for some reason.

@AndurilObserver Good ol bangin in sangin. Not a great place to be playing in the dirt though!

Love to see my fellow Marines demonstrating the Corps' ability to adapt to new technology and weapons systems. Despite being the smallest and least funded branch, they consistently lead from the front in the evolution of modern warfare.

Any evidence recovered is not considered properly marked unless it has one or more dicks drawn on it. This was taken around the time we received directives to stop trying to recover evidence from these devices unless circumstances dictated that it was absolutely worth the effort. The forensic labs were overloaded and totally backlogged and EOD Techs regularly getting hurt/killed working on devices that (usually) could have been just blown-in-place (BIP). We tended to keep trying to recover evidence anyway. I think we all thought, somewhere in the back of our brains, that we would crack the IED network with evidence we recovered. Sometimes we did make a dent, but it didn't really ever work out the way we hoped it would.

@AndurilObserver I onow you EOD guys know which fork goes with each course at dinner and all but did you draw a dick on it