kimbosbread

Sealift + airlift + logistics personnel + training + ISR assets in air and space + missles like ATACMS. It's a safe assumption that a lot of this equipment would get blown or used very quickly, and any of it in Ukraine would be hindering an effort in the Pacific.

You don't need a gun at all. Just go Phantasm style and put a drill with an electric motor on the front of the drone and fly straight into the target's forehead and drill baby drill. You can cobble together incredibly good face/head tracking software that will run on minimal hardware.

Me too. I doubt we'll have to wait long though. On a similar subject, I wonder if there's gonna be a rapid evolution in sniper rifles soon, where the spotter + sniper converges with kamikaze drones.

The US should show Europe who is boss and admit Ukraine + Moldova + Georgia to NAFTA. Forget the EU, our club is more awesome plus we have poutine and tacos.

I suspect standarization is key. If it's a built in battery at the factory, at least same charging interface. If it's a fuel, same fuel. I wonder if otto fuel (from torpedoes) would be useful, especially for single use drones? Yeah if your per drone cost is $1000, but you have a good kill/useful intelligence ratio every time you consume a drone, maybe it's not even a driving consideration. So for example if you decide to deploy your drone swarm in area X, do you consider that to be the same as CAS where let's say you've burned $300k of fuel, maintenance and airframe (or the equivalent in humvee hours and associated back surgery/physio for the crew), so you just say whatever, all the drones are considered consumed? It's really about the choosing which of those conflicting requirements are the most important. Do you care about reuse? For micro cruise missiles, I suspect not. It's basically a flying smart mine that can coordinate with all the other smart flying mines without phoning home. For Mr Robotank, I suspect reuse is much more important, as well as maintability and fitting into the supply/maintenance organization with wheels, ammunition, fuel, poorly applied reactive armor tiles etc.

Rechargeable means only using half of your battery charge every flight, unless you are find junking that battery in 100-200 use cycles. If you have small drones (fixed wing, at least not power hungry quadcopters) with longer ranges they can make their way much further back to find a recharge pad (let's say no battery swap) or for a refuel. Nice thing about sealed rechargeable battery is you could just slap an induction pad on it and although charging is slower, there is no port for dirt to get into. It also means you could just drop a bunch of passive self contained recharge pads on the battlefield that would be good for 100 charges and you consider them a consumable (a nasty, toxic lion one though). Swapping batteries is much faster if you can figure out a good autonomous way to do it.

As a relatively uninformed software engineer, I think this is where things will go. We've been able to build small (10cm long) fixed wing drones that can fly for an hour at 100km per hour for 24+ years with a camera (at least when I was visiting the MIT aerospace labs in 1999). You can build something like this now for very cheap, and much improved performance just on the hardware side. Let's say these are electric drones, so trade endurance for no need to truck fuel around and simpler motors, you could have thousands of these things in the air, with 25g of high explosive in the nose. Combine that with modern ML/image recognition software + GPS running on hardware that fits into a tiny package AND is cheap and easy to build, as in $100 or less cheap. Effectively this is a smart cruise missile scaled down to kill an individual soldier. Let's say Ukraine could have a ten thousand semi autonomous drones roaming 20km behind Russian lines looking for targets for several hours at a time; would you even need mass if it was cheap to dedicate a drone to kill each enemy soldier? If the defences are cheap enough to scale with the number of drones, maybe. Lasers seem reasonable, but they require a lot of power and cooling, especially for longer range. As soon as a laser is identifier, something specialized can be sent after the laser. The more insiduous thing with drones and small computers is there is no reason the drones can't be used to execute flexible tactics. For example, we've identified a laser on a bmp-9000, and we estimate it can only fire 20 times per minute, so let's swarm it with 25 drones at once, or attack from multiple directions at once with 5 drones.