No, The New Kalashnikov Is Not a Weapon – A rebuttal to Gerald Mako's @smallwars opinion. Cheap commercial drones fitted with readily available guidance software have collapsed the cost of precision strikes. Ukraine has demonstrated the method at scale: mass-produced frames, improvised warheads, and increasingly autonomous terminal guidance have given states and non-state actors capabilities once reserved for conventional military forces. The warfare equation has changed. That much is not disputed. What does not logically follow is the policy package offered to reverse it. Gerald Mako's argument for tighter controls on hardware that enables onboard autonomy, enforceable obligations on manufacturers whose platforms are routinely militarized, and a minilateral coalition of component producers treats the commercial dual-use ecosystem as if it were a controllable arsenal. It absolutely is not. The systems in question began as hobby, racing, photography, and agricultural products and became weapons only after conversion. The proposed remedies – manufacturer due diligence, conversion-resistant firmware, and supply denial of enabling components – do not survive contact with how these systems are actually built and fielded. The AK-47 analogy fails as an analogy. The Kalashnikov was purpose-designed as a military weapon and then deliberately proliferated as a geopolitical instrument: cheap, rugged, and simple to operate by design. Small UAVs and FPV platforms followed the opposite path. They began as commercial products before being adapted, often in small shops and wartime factories, into delivery vehicles. The distinction is acknowledged in passing and then ignored. A weapon that begins as a weapon can be regulated as one. A commercial airframe that becomes a weapon only after a user straps on a warhead or swaps the flight controller presents a fundamentally different regulatory problem. Imposing strict due-diligence duties, secure-boot mandates, and market-access conditions on commercial manufacturers because downstream actors convert their products is the gun-control argument applied to dual-use electronics. A manufacturer cannot reasonably be held responsible for every illicit modification any more than a tire company can be held liable for a drunk-driving manslaughter. The article concedes that no firm can be responsible for every conversion, then recommends treating refusal to adopt “reasonable” anti-militarization features as grounds for exclusion from Western markets. That substitutes upstream obligations for downstream enforcement. Such measures may deter some casual misuse and impose compliance costs. They will not stop a knowledgeable bad guy. The practical failures are straightforward. Software is already open and mirrored. Jailbreaks, forks, and companion computers have been routine in the hobby community for years and are now industrialized. Physical conversion does not require defeating firmware. A warhead and basic fuze can be zip-tied or strapped to a commercial platform. Motors, props, batteries, and boards can be stripped from a “conversion-resistant” airframe and reused on an open stack. Middle markets and gray-market channels for commodity components persist as a structural feature of global trade, not a temporary loophole. These realities make conversion resistance a weak foundation for policy. Local fabrication has already collapsed the premise that the original commercial stack is an indispensable chokepoint. Ukraine’s wartime industry is the existence proof. Firms there now machine motor housings, wind stators, fabricate or assemble flight controllers, ESCs, and initiation boards, and produce frames and props at meaningful volume. Localization of 70–90 percent or more is documented across multiple producers, with isolated examples of near-full domestic component sets for FPV systems. Residual dependencies remain – high-grade magnets, battery cells at volume, certain specialized sensors, higher-end edge chips – but they are narrower than a comprehensive supply-denial regime assumes. Once process knowledge and modest equipment exist, a determined actor does not need the original manufacturer for an effective short-to-medium-range system. Targeting commercial makers or the commodity layer cannot be the primary line of effort. The coherent response is therefore actor-centric. If the concern is bad actors using drones, policy should target those actors and the systems they field. That means intelligence and interdiction against the specific groups, workshops, conversion networks, and deployment chains that assemble, arm, and employ them. It means layered detection and defeat once the systems are in the air. It means narrow restrictions only where they can be tied to identified diversion networks and the few high-end items that still materially improve particular programs. It does not mean spreading friction across legitimate commercial markets, component producers, and software repositories in the hope of raising costs for actors who have already demonstrated their ability to substitute and fabricate around those frictions. Two clocks are running, but not the clocks described. One is the conversion cycle on commercial frames. The other is the localization and substitution cycle that has already outpaced upstream controls. Geneva processes and minilateral component coalitions will not stop what small shops and motivated actors can already build. The practical task is to target the people who turn available technology into weapons, not the manufacturers and markets that produce the technology.

The AUTOWAR25 Morning Edition is out and the links are active. We have a few small formatting tasks to resolve, but everything is live.

The darkness that is Marxism in power will never allow that power to simply be voted away from it. It is a venom for which there is no antidote. No antidote but one.

Consider the reality that there is a huge market need for cheap, mass-produced, disposable, jet engines.

Most American bases used to generate their own power and run their own grids. Connecting everything to the civilian system was a cost-and-convenience choice, not a readiness standard. FORTRESS is less “preparing for the coming storm” and more correcting a decades-long readiness shortfall that left critical missions dependent on infrastructure the military doesn’t control.

🤌

The communist dwarf does what the communist dwarf always does, demands communism in America.

As I said yesterday, the post below is my favorite sort of post on X. I’m going to apply it to my own domain of study. Time is the one resource in war that cannot be recovered. Lost ground can be retaken and lost resources replaced; a lost hour is gone for both sides forever. Duration, frequency, sequence, and opportunity must be managed together – space is only where the fight happens, time is how it is won. Time is the primary constraint. Yet not every minute carries the same weight. The decisive resource is what I call the cognitive minute: the finite human attention that still sets usable tempo, sequence, and the recognition of fleeting opportunity. When that attention is spent on repetitive coordination, routine state-keeping, or manually sequencing already-understood actions, the force is consuming its cognitive priority on menial continuity. Opportunities close and the tempo of the fight slows to the rate of human attention. Automation does not solve this challenge. In many cases it creates additional cognitive debt for the commander and operator – now they are coordinating at a distance while trying to keep up with the speed of electronic process. It is an impossible ask. But this is also why AI and autonomy have become the central practical solution – computational challenges require computational solutions. They are the tools available for moving continuity of already-authorized action off continuous human presence, so that the cognitive minute is no longer spent on menial coordination. The objective is not more autonomous platforms for their own sake. It is to keep the larger operational process coherent without requiring humans to maintain every transition. Only then does additional speed free, rather than further consume, the human attention that time itself requires. That effort requires care. If autonomy is applied by making every endpoint more capable while leaving the work of keeping the larger operation coherent with people, the volume of menial coordination simply increases. Cognitive priority is preserved only when routine continuity is removed from the human domain while authorship of intent, constraints, and the decision to change the plan remains there. Without that boundary, the current focus on AI and autonomy risks accelerating the consumption of the resource it is supposed to protect.

Prediction markets on Congress races are actually worse than polls because they’re easy to manipulate — there’s no volume. A single transaction can drastically move a percentage and keep the number mispriced for weeks. And yes, the polls are garbage too.

Worth a listen. Never forget.

Haití says It’s overwhelmed by… Haitians. Not wrong.

Why is my feed totally infected with crap movie clips?