This is my third piece this week on how the world is changing and why. The first handled the geopolitical, the second the military tech at this moment and how that is making empire difficult.
This one is about the future.
There’s going to be a period of war which is all about autonomous robots. Drones, missiles, robodogs with guns, tiny swarms, etc…
Humans are a stupid and inefficient way to apply force: most of the human body is not designed for combat: we are slow, clumsy and easily damaged and destroyed compared to what we can build.
As the cost of autonomous robots (and they will be autonomous because remote control is a weakness) continues to plummet and as the knowledge of how to build them spreads, they will replace humans on the front lines. Humans will be victims, but not primary combatants.
At the state level this means that states which can produce the most robots will win: the robots will be expendable and used in vast numbers. The chain of resources to manufacturing and the ability protect that chain will be what matters.
For smaller groups, robots will offer cheap violence against soft targets (and sometimes hard targets.) A militia can be people who build drones then use them to attack a governor or an activist they hate.
Let’s give one concrete example. Say it’s twenty years from now, you’re China and some piss–ant country like Yemen is causing problems hitting your ships with drones and missiles. You warn them and they don’t stop.
Fine. Release a few million autonomous hunter-killer drones. They will crawl over every single inch of land, not even in the mountains will it be possible to hide. No matter how many robots Yemen has, you’re China, you have magnitudes more. You can’t lose.
In time there will be, as the gamers say, a “meta”–we’ll figure out how autonomous robots work, and how to fight with them and defeat them and so on. But during the adoption period (and remember, that period is usually 30-40 years and sometimes longer) those who figure out how to use robots best will punch far higher than their apparent weight, and if anyone can obtain a monopoly on some for of advanced weaponized robot which is effective (like European ironclads when no one else had any), well, they will do very well and may be able to parlay that into a long period of dominance.
Don’t be sure you know exactly how this will play out. For example, a decentralized model where every citizen builds and contributes drones may turn out to be very strong versus a centralized model. Or it may not. We don’t know yet.
But the time of meat as the right way to fight is coming to an end.
(Or has it? We’ll come to that in the next article in this series.)
Our fundraiser ends Friday. We’re about $400 from the final goal, which is an article on the Medieval credentials crisis.
different clue
I suppose the target of the multi-million robot swarm will try to hack into the central robot swarm brain and redirect the swarm to return to base and attack the senders.
In this scenario, I suppose the Yemenis or friends of the Yemenis could also try sending a counterswarm to exterminate the senders of the robot swarm. In this scenario the Yemenis would hope that the swarm sender would not live to appreciate and enjoy the victory.
Or maybe the Yemenis will try to counterplay on a different battlefield and send swarms of Plague releasers or SuperCovid releasers or Some-New-Disease releasers to the homeland of the robot warm senders in hopes of giving every single one of the people involved a lethal new disease.
If the attacker can out-robo you, perhaps you can out-bio the attacker in return.
Soredemos
I’m going to take a contrarian tone and say that none of that happens because there are hard practical limits to what ‘AI’ (actually just dumb automated algorithms) can do, and will ever be able to do. Sitting down and spending time interacting with various flavors of ‘AI’, I’ve been utterly unimpressed with how any of them operate. Garbage in, garbage out, easy to spot, incredibly easy to break. And above all, not an ounce of anything that passes for thought going on. I suspect humans would more than stand a chance evading or defeating dumbass glorified ChatGPT droids.
That’s assuming, by the way, that anything like a million mini-bot swarm were ever be made. I suspect there will end up being hard manufacturing or other logistics limits to building such an army. I know, I know, self-replicating nanomachines or whatever. Watch as none of that happens either at scale or with any degree of sophistication.
Cool ideas for a scifi story though. The killer robot swarm is a key part of the backstory to the Horizon: Zero Dawn games.
mago
Yemen is a piss ant country?
Probably.
The piss ant is a small red ant that got its name from bitting the human nut sack.
That’s from my store of Idaho folklore.
It fits.
Oakchair
“I suspect humans would more than stand a chance evading or defeating dumbass glorified ChatGPT droids.” —Soredemos
When it comes to humans of yesteryear and many of today I suspect this as well. However, various trends makes this less true every single year. In the last 50 years Autism rates in children have gone from 1 in 10,000 to 1 in 36 (USA). Not only is that just one of the numerous negative health trends, but it is the the effects seen only at the end of the bell curve. We’re 4+ decades into a mass poisoning event similar to how Rome and the Mayan’s used lead and mercury piping and paint.
There is a high chance that the manufacturing limit for a droid army will be that society lacks the human capacity, and resources to make it because everyone will either be to poisoned, or to busy taking care of all the poisoned people.
Purple Library Guy
I dunno. I partially buy this. There will be lots of drones, they will be important in war, some of them will be autonomous.
But, “AI” is not as artificially intelligent as most people think it is, and the “AI” that we have is a fairly mature technology. Yes, there will be autonomous killer robot/drones. But, I don’t think they will be very smart. Not “not very smart” in the way a dull human is not very smart, though. Rather, they’ll be really amazing at whatever they were trained for, but if you depart from that script, by say dressing up as a smurf, or hanging upside down or whatever, they won’t know what you are. War will get sillier as word gets around about how to get around drones’ machine learning. And you know, you can train them for any given deviation, but the problem is that the current “AI” paradigm is not artificial intelligence in any sense resembling what we think of as “general intelligence”. These things do not understand what they are doing, they do not have any definition or concept of what the problem space is so that they can fit different things into it. What they have is success at training whatever situations their trainers threw at them–and they can be very good at whatever that is, but it does not get them anything outside that. So if people start acting weird in one way, and the next generation gets trained to deal with that, they will still not have a concept of “acting weird”; train them to notice people disguising themselves as deciduous trees, and they will probably still not notice people dressed as pine trees, let alone ferns or mushrooms.
Second, there will be swarms of cheap drones, but they will mostly be normal sized, because tiny ones can’t carry enough fuel to go very far or stay active very long, or enough payload to do very much. There will be little scouty drones with sensors in specialized roles–say maybe you’ve got your big-ish robot that gets fairly close to where there are probably some enemies and sends out some tiny drones it controls to report back on just where they are so it can go kill them. But millions of little drones all acting autonomously? Nah.
Purple Library Guy
On the tiny drones: Consider that evolution has almost universally given us apex predators that are big animals rather than swarms of mouse-sized ones. Sure, there’s piranhas, but even there–I’ve heard of piranhas stripping capybaras to the bones, but I never heard of them bothering a caiman.
Allison
Ian, I’m thinking it may be time to get that commenter-help thread going again. You know, the one where “different clue” aka “i just don’t like the gravy” posts all manner of largely “low-tech” alternatives that may or may not be helpful to others, while Astrid posts about bread makers and other devices that commenters may or may not be able to afford.
How else will we survive the drone apocalypse?
Tallifer
I am reminded of the shock round the world when Azerbeijan overwhelmed the Armenian army with drones in 2020.
StewartM
I’ll just repeat my previous comment:
Who will be controlling these drones? Will they be controlled by humans via signals? Those signals can be jammed.
If they’re AI-controlled, then I forsee that too being defeated. Unlike AI-driven autos, where the terrain and road and other drivers aren’t deliberating trying to hit you, in a battlefield the other side will learn to use drones’ own algorithms against them.
How does this reference Clausewitz? After all, a war is fought to obtain some political aim, not to just obliterate people and structures. Which is why I find it difficult to imagine the end of infantry. You have to occupy and control the terrain you occupy.
Unless, of course, this includes a even more dystopic RoboCop future.
GrimJim
“After all, a war is fought to obtain some political aim, not to just obliterate people and structures.”
We are reaching the point once again where extermination of the enemy and all they have wrought is the purpose of war. If MAGAts could send nanoswarms against “Libruls” they would. And not just here, Reich Wing thought is coming into practice everywhere…
Plus, in the future, it will be a matter of survival, to exterminate your enemy, not only before he exterminates you, but also for your immediate need for their resources as your quickly drown or burn…
mago
Hello Allison. Astrid and Hugh checked out from this site a while back as have a few others.
Blizzard of Oz how are you?
I fear Che Pasa may have shuffled off this mortal coil, or maybe he objected to Ian’s stance on Israeli apartheid.
Don’t know.
Like Astrid he was a thoughtful and insightful commentator.
Maybe it’s like the spud who had a crush on Walter Cronkite and her mother said, “forget about him, he’s just a common tater. “
Never mind. . .
Carborundum
Ah, swarm warfare. Almost up there in the same category as fusion. Been 5 years out for the last 20.
I exaggerate a little for humour – there’s some merit to the strategy, though it will play out vastly differently from the Cyberpunk 2077 dystopia that seems to seize everyone’s imagination. Think more along the lines of the bastard child of ubiquitous ISTAR and distributed, tech-enabled kill chains. Perhaps I’ve been around this too long, but I tend to think only people who haven’t seen the two way range up close want autonomous systems – folks who have experienced the local universe’s propensity to become pear-shaped for entirely unpredictable reasons are much more interested in using the tech as amplification.
In terms of this specific hypothetical, I would note that while there’s been lots of interesting work done in aerospace and maritime domains, the terrestrial stuff has pretty much sucked. Way more challenging environment for these systems.
Allison
mago,
And what of Trinity? Gone as well.
I mostly miss DMH:
https://www.ianwelsh.net/open-thread-206/#comment-146268