What a 20-year-lead in military tech might look like

I’ve spent way too much time spec­u­lat­ing about near-fu­ture mil­i­tary tech­nol­ogy. Here is a list of tech­nolo­gies I think are likely to be im­por­tant to a con­ven­tional war in, say, 2040.

If you think this list shouldn’t be taken se­ri­ously be­cause I don’t have offi­cial ex­per­tise, you are prob­a­bly right. I think that if I be­came an ex­pert I’d prob­a­bly end up delet­ing a third of the things on this list and adding a similar num­ber of new things. (I did run this draft past three peo­ple with rele­vant ex­per­tise, though!) I look for­ward to cri­tiques in the com­ments.

If you are won­der­ing why these spec­u­la­tions are rele­vant to LessWrong… well, it was re­quested, and also it’s rele­vant to a point I want to make later about AI. You see, the tech­nolo­gies on this list could prob­a­bly be quickly achieved with the help of ad­vanced-but-not-godlike-AI. The gap be­tween these tech­nolo­gies and pre­sent-day mil­i­tary tech is an ex­am­ple of a “small” gap; a “merely” 20-year-by-hu­man-stan­dards gap. So, for ex­am­ple, sup­pose some cor­po­ra­tion or na­tion gets AI sys­tems which are good but not godlike at de­sign­ing new tech­nolo­gies. They vastly speed up the pro­cess but don’t do any­thing hu­man en­g­ineers couldn’t do in 20 years. Or, sup­pose there are lots of AI sys­tems in the world but one fac­tion has sys­tems which are the equiv­a­lent of 20-years-by-hu­man-stan­dards ahead. In ei­ther case, I think, the gap would trans­late to an ex­tremely large ad­van­tage, pos­si­bly enough to en­able a small group to take over the world, con­quis­ta­dor-style.

(To be clear, while I do think AI takeover is prob­a­ble, I think per­sua­sion/​poli­tics/​ide­ol­ogy is by far the most likely method.)

Bet­ter com­mand-and-con­trol ca­pa­bil­ities (incl. Cy­ber and in­tel)

Com­mand & con­trol tech­nolo­gies are very im­por­tant. Wars are won or lost on the ba­sis of how well each side is able to ob­serve, pre­dict, plan, or­ga­nize, and out­ma­neu­ver the other. I think the ad­vance­ments in com­mand and con­trol tech that are likely to hap­pen in the next 20 years are more im­por­tant than ev­ery­thing else on this list com­bined. In case this isn’t ob­vi­ous, I’ll say a few things be­low by way of jus­tifi­ca­tion.

Long-ranged guided weaponry already makes it easy to blow up tar­gets if you know where they will be when your at­tack ar­rives. Win­ning is all about find­ing the en­emy be­fore they find you. Now more than ever, be­ing able to quickly gather loads of in­for­ma­tion and in­te­grate it and com­mu­ni­cate the re­sult­ing or­ders to your forces is su­per im­por­tant, as is be­ing able to dis­rupt your en­e­mies’ at­tempts to do the same.

There’s lots of room for new tech­nolo­gies in this space, or much more ad­vanced ver­sions of already ex­ist­ing tech: Satel­lites, spy drones, ad­vanced sen­sors of var­i­ous kinds, more ro­bust com­mu­ni­ca­tion net­works (e.g. cloud com­put­ing), AI to in­te­grate all the data, draw some in­ter­est­ing con­clu­sions, and in­stantly make some of the rele­vant de­ci­sions. Already the US mil­i­tary col­lects way too much data for teams of hu­man an­a­lysts to ex­am­ine, and any­how hu­man an­a­lysts take too much time. Bet­ter to have a com­puter say “Odds of im­mi­nent at­tack in sec­tor 7 have spiked to 11%” as soon it no­tices some un­usual pat­terns. (A ver­sion of this might be use­ful at the tac­ti­cal level, e.g. “The sniper who shot that bul­let is prob­a­bly in this win­dow; the mor­tar shell was prob­a­bly fired from the alley at co­or­di­nates XY.”)

What would it look like to have a sig­nifi­cant ad­van­tage in C&C+in­tel+cy­ber tech? The en­emy would be swing­ing blind punches, firing mis­siles at lo­ca­tions where your troops aren’t, or march­ing for­wards in the hopes of mak­ing con­tact while your forces quietly dodge around them or set up an am­bush. Your air­craft would fly low over their coun­try, slip­ping through gaps in air defenses ex­posed by your satel­lites, de­stroy­ing key nodes of com­mu­ni­ca­tion and trans­port. In re­gions you oc­cupy, in­sur­gency would be im­pos­si­ble, as ev­ery­thing that moves (at least in im­por­tant re­gions) would be tracked by your cam­eras. In re­gions the en­emy oc­cu­pies, your in­sur­gents (prob­a­bly lo­cal al­lies sup­plied with weapons and guidance) would be a real threat. Mean­while, you might have all sorts of elec­tronic spies in their net­work, due to su­pe­rior cy­ber ca­pa­bil­ities. You might even be able to shut down some of their sys­tems at cru­cial mo­ments.

Per­sua­sion tools might also be pos­si­ble. Pro­pa­ganda is already a thing, but per­son­al­ized pro­pa­ganda pow­ered by big data and ma­chine learn­ing will be much more po­tent. More­over, it is use­ful in peace­time also, and use­ful to many non-mil­i­tary ac­tors, so it will be heav­ily in­vested in and rapidly de­vel­oped. This is an­other rea­son why hav­ing an ad­van­tage in this sort of tech will help sup­press en­emy in­sur­gen­cies (and su­per­charge your own). It might also help with “turn­ing” key en­emy per­son­nel (such as a cap­tured leader, or even a leader you have com­mu­ni­ca­tion ac­cess to via a spy or hack or par­lay). At the very least it will help keep your own troops’ morale and loy­alty high, which has always been and still is ex­tremely im­por­tant.


Aim­bots are guns that aim and fire au­to­mat­i­cally. A sim­ple ver­sion would be a cam­era and com­puter hooked up to the trig­ger of an in­fantry rifle, such that the rifle fires when the com­puter calcu­lates that the bul­let would hit the tar­get. The soft­ware in­volved would be a bit of image recog­ni­tion, maybe a bit of physics, etc. The user would point the gun at the tar­get and the bul­let would fire at pre­cisely the right mo­ment, in­creas­ing ac­cu­racy while de­creas­ing aiming time. Th­ese aim­bots already ex­ist and are be­ing tested by the US mil­i­tary.

A more com­plex ver­sion would have some way of point­ing the gun. For ex­am­ple, a can­ister of com­pressed gas as an un­der-bar­rel at­tach­ment, with mul­ti­ple noz­zles point­ing in differ­ent di­rec­tions. The com­puter se­lec­tively opens the noz­zles for split-sec­onds, emit­ting jets of gas that jerk the bar­rel in such a way as to line it up with the tar­get, at least mo­men­tar­ily, dur­ing which the com­puter fires the bul­let. Per­haps the cam­era would also be con­nected to a cell phone screen, so that the sol­dier hold­ing the rifle can stick it out from be­hind a wall while his buddy looks through the cam­era screen and clicks on tar­gets.

Any­one who has played a first-per­son shooter with an auto-aim hack or abil­ity will already think this is pow­er­ful. A rifle at­tach­ment that quick­ens your re­ac­tion times and makes al­most ev­ery shot hit? Amaz­ing! How­ever on a real bat­tlefield the effect would be much more pow­er­ful than it is in video games. Here is why:

The vast ma­jor­ity of bul­lets fired in war are not aimed, but sprayed in the gen­eral di­rec­tion of the en­emy. This is called “sup­press­ing fire;” it keeps the en­emy pinned in place and pre­vents them from shoot­ing back at you. Sup­press­ing fire is mostly not a phe­nomenon in video games be­cause play­ers don’t die in real life if they get hit by a sin­gle video game bul­let. In fact in most games a sin­gle bul­let doesn’t even kill your video-game char­ac­ter, and if your char­ac­ter does die he sim­ply respawns. In real war, how­ever, sup­press­ing fire is the pri­mary way to deal with en­emy in­fantry. They are usu­ally in cover, so you can’t just shoot them; how­ever, you can sup­press them, so that your own forces have free­dom to ma­neu­ver safely. Then, your own forces can flank the en­emy po­si­tion, or close with it and lob grenades, or what­ever. Of course, they’ll be try­ing to pre­vent you from do­ing this by sup­press­ing you; you “win the fire­fight” when they are sup­pressed and you are not.

Now sup­pose that one side has aim­bots on their rifles. They can now shoot their rifles with­out pok­ing their heads out of cover (the point-and-click method men­tioned above). This means they can provide sup­press­ing fire with­out en­dan­ger­ing them­selves. This means you can’t win the fire­fight against them, at least not in the nor­mal sense. Mean­while, their shots are way more ac­cu­rate, and their re­ac­tion speed is in­stan­ta­neous—mean­ing they don’t have to spray bul­lets in your gen­eral di­rec­tion, they can just wait, silently, and shoot a sin­gle bul­let at you when you poke your head around the cor­ner. The bul­let will hit you be­fore you can re­act to what­ever it is you’ve seen. What this means is that (1) your forces will be sup­pressed ex­tremely quickly, even if you out­num­ber the en­emy and start shoot­ing at them be­fore they start shoot­ing at you, (2) you’ll lose a few men be­fore you re­al­ize what is hap­pen­ing and hun­ker down, (3) their am­mu­ni­tion will last much longer be­cause they make ev­ery bul­let count, and (4) af­ter the en­emy leaves you might still be sup­pressed be­cause you won’t know when they are gone be­cause you are afraid to look. (With or­di­nary hu­man-caused sup­press­ing fire, the shots are be­ing sprayed at you and are un­likely to hit, so if you are brave you can peek for a sec­ond at rel­a­tively small risk. Prob­lem #4 can be solved to some ex­tent by coun­ter­mea­sures like us­ing your own cell phone or a mir­ror to peek around the cor­ner, as­sum­ing their AI isn’t good enough to shoot it too.)

What this means is that an in­fantry force equipped with aim­bots would be not only much more deadly, but also much harder to kill or stop. It would move through the bat­tlefield like a basilisk, par­a­lyz­ing and/​or de­stroy­ing any­one who comes within line of sight. Of course, if both sides have aim­bots, the fight would be much more fair.

Aim­bots would help enor­mously against drones. See next sec­tion. Aim­bots might also lower the amount of train­ing needed be­fore a re­cruit be­comes bet­ter than use­less.

In­stead of an at­tach­ment to an in­fantry rifle, an aim­bot could be im­ple­mented as a portable au­to­tur­ret. It’s on a tri­pod or some­thing, you plop it on the ground and it can aim and fire au­to­mat­i­cally, au­tonomous or re­mote-con­trol­led. I think this would be less effi­cient than the rifle im­ple­men­ta­tion but it would have its ad­van­tages too. For ex­am­ple, you could at­tach them to your ve­hi­cles. This already ex­ists and has been quite effec­tive, see e.g. Tro­phy.

Bat­tle bots

Th­ese are ma­chines which pro­pel them­selves around the bat­tlefield, iden­ti­fy­ing and at­tack­ing tar­gets. Early ver­sions are mostly re­mote-con­trol­led by hu­mans, later ver­sions are more and more au­tonomous. There are many va­ri­eties, and re­lated con­cepts. The core idea is that bat­tle bots are cheaper than hu­man sol­diers; for the price, has­sle, and trans­porta­tion cost of one hu­man sol­dier you can de­liver dozens if not hun­dreds or even thou­sands of bat­tle bots to the field. There are other ad­van­tages too. For ex­am­ple, they don’t have morale prob­lems, and they always obey or­ders. On the other hand, they lack ini­ti­a­tive and cre­ativity and lose more of their effec­tive­ness when com­mu­ni­ca­tion with com­mand is cut off. But these dis­ad­van­tages will lessen as the tech­nol­ogy im­proves.

Drone swarms are per­haps the most im­por­tant kind of bat­tle bot. Kamikaze drones fly into the en­emy and ex­plode. Gun drones shoot bul­lets. Utility drones can trans­port cargo, open doors, recharge other drones, drag huge nets through the sky to catch other drones, scout with big­ger cam­eras and com­put­ers, drop bombs, and many more things, de­pend­ing on how they are equipped. (All of the above would come as mod­ules that can be at­tached or de­tached eas­ily by a hu­man han­dler). I’m not sure whether kamikaze or gun drones would be more preva­lent. Cur­rently kamikaze drones are. (Note: I the pre­vi­ous links I found af­ter writ­ing this. It’s nice to see my pre­dic­tions con­firmed.)

Drone swarms, since they can fly, can close with the en­emy very quickly. Since they are cheap and small, you can have very large swarms and trans­port them to the bat­tlefield eas­ily. (You can even just drop them out the back of an air­plane!) Say a drone car­ry­ing an ex­plo­sive charge the size of a grenade costs $10,000 — a con­ser­va­tive es­ti­mate con­sid­er­ing that the com­puter on the drone is prob­a­bly the most ex­pen­sive com­po­nent, and wouldn’t be more than $2,000. Then you can drop 1,000 of them on the bat­tlefield for about the price of a tank (crew not in­cluded). Or 200 for the price of a cruise mis­sile. Swarms this big would be very hard to kill. They are just a bunch of dots in the sky mov­ing very fast and maybe wig­gling er­rat­i­cally. With aim­bots you could shoot them down, but even an au­to­tur­ret would prob­a­bly only be able to take out 10 or so be­fore they closed in on it and blew it up. And de­pend­ing on how dense the ter­rain was, maybe that num­ber would be more like 1. So the swarm would just steam­roll over ev­ery­thing in its path un­til it ran out of bat­tery, or was stopped by a force of aim­bots or bat­tle­bots of similar size. (Or­di­nary hu­man in­fantry would be much less effec­tive. Even if you have good aim and good re­flexes, you prob­a­bly won’t be able to shoot more than 1 kamikaze drone out of the sky be­fore the swarm is upon you.) Hid­ing in build­ings or un­der cars or un­der camoflauge would be an effec­tive tac­tic against drone swarms, but not su­per effec­tive — af­ter all, while you are hid­ing you can’t con­tribute to win­ning the bat­tle, and more­over the drones can just land like lo­custs, con­serv­ing bat­tery and wait­ing for tar­gets to come out of hid­ing. (This is what makes drone swarms bet­ter than ar­tillery bar­rages or mas­sive airstrikes. That, and the abil­ity to dis­t­in­guish friend from foe from civilian. And the abil­ity to redi­rect to some­where else, or even come home, if the en­emy turns out to no longer be in the tar­get lo­ca­tion. And the abil­ity to sweep a broad area look­ing for en­e­mies.) More­over, some kinds of drones (quad­copters) might be able to fol­low you in­doors. See the Slaugh­ter­bots video, which has some large kamikaze drones for tak­ing out doors and win­dows, and then small kamikaze drones for kil­ling hu­mans. You could hang beads and cloth in door­ways and hal­lways to im­pede drone move­ment, but util­ity drones with at­tach­ments could get around that.

So far we’ve been dis­cussing swarms of small drones. But swarms of big drones will hap­pen too. Air-to-air com­bat in the age of au­tonomous ve­hi­cles is very much a num­bers thing. How many mis­siles can you bring to the bat­tle? How many dis­tinct air­craft do you have, so that the en­emy needs to bring at least that many mis­siles? How much range do your mis­siles have? We already have big Preda­tor and Reaper drones, fit­ted with var­i­ous kinds of mis­siles. Just make loads more of them. In world war 2 the biggest air raids had thou­sands of planes. Nowa­days we could prob­a­bly make at least that many big drones, and prob­a­bly many more. (Spot check: Preda­tor costs $4M, so the Pen­tagon could eas­ily af­ford to buy 10,000 of them per year, even with­out ac­count­ing for economies of scale.) “Air raid” will prob­a­bly be a mis­nomer; “Air oc­cu­pa­tion” more like it. A storm­front of drones smashes into en­emy ter­ri­tory, ex­chang­ing mis­siles with en­emy air­craft and anti-air defenses, los­ing many drones in the pro­cess but not nearly enough. Drones that get low on fuel or ammo go home, and are re­placed by re­in­force­ments. Enemy ter­ri­tory is par­a­lyzed due to con­stant cir­cling drones over­head, ready to drop mis­siles on any sus­pi­cious ac­tivity.

Mini­tanks are au­to­tur­rets with wheels. Maybe they have some ar­mor too, maybe not. They are much less mo­bile than drones, even less mo­bile than hu­man in­fantry. Why build them? Well, if you need lots of auto-aimed guns in a cer­tain lo­ca­tion, maybe you don’t have enough hu­man legs to carry them. Or maybe your hu­mans are too valuable and frag­ile. Why not just build drones in­stead? Well, maybe one mini­tank can shoot fast enough and is ar­mored enough to be worth 10 drones in most fights, but costs only as much as 3. Ground travel is more en­ergy-effi­cient than air, and it can af­ford to have ar­mor. Mul­ti­ple mil­i­taries are already ex­per­i­ment­ing with these.

Drone sub­marines are to naval war­fare what large drones in the sky are to aerial war­fare. They’ll be much smaller and cheaper than or­di­nary sub­marines; maybe they won’t even launch tor­pe­dos, maybe they’ll be kamikazes in­stead. Swarms of them will block­ade en­emy coastlines, ruin en­emy trade routes, etc. The US mil­i­tary is already ex­per­i­ment­ing with this.

Bal­loon bombs are a spe­cial kind of drone swarm. In World War Two, Ja­pan made about 10,000 bal­loons that car­ried bombs, and sent them up into the jet stream to drift over the USA and drop their bombs. At the time they didn’t have any sort of abil­ity to aim the bombs or the bal­loons, so very lit­tle dam­age was done. How­ever, pro­ject Loon has demon­strated that mod­ern tech­nol­ogy can steer bal­loons pretty much any­where you want them to go. It just takes a while for them to get there. Once there, the bombs they drop can be smart bombs, that iden­tify tar­gets us­ing cam­eras and GPS and glide right to them. A glide bomb could hit a tar­get within about 70 miles of the bal­loon it drops from. They could also carry air-to-air mis­siles. Bal­loon bombs are like large drone swarms, ex­cept that they are much much cheaper. It might even cost more to shoot them down than to make them! If Ja­pan could make 10,000 eighty years ago with a tiny frac­tion of their war effort, a mod­ern na­tion with a big­ger econ­omy could be mo­ti­vated to make mil­lions.

Drone car­ri­ers are to tanks and ar­tillery what air­craft car­ri­ers are to bat­tle­ships. A drone car­rier could be a civilian pickup truck or trailer truck, with crates of drones, maybe a gen­er­a­tor, loads of charg­ing ca­bles and spare bat­ter­ies, and a few peo­ple in the back. Some­one drives, maybe some­one else is on their lap­top giv­ing com­mands to the drones, and the peo­ple in the back un­pack drones from crates and launch them on strikes and then grab re­turn­ing drones out of the air and swap out their bat­ter­ies. By mil­i­tary stan­dards civilian trucks are prac­ti­cally free, so a drone car­rier is ba­si­cally as ex­pen­sive as the peo­ple and drones it car­ries. And the peo­ple need not be highly skil­led. And the drones are cheap. Just as bat­tle­ships be­came ob­so­lete, so too will tanks and maybe even ar­tillery be­come ob­so­lete. Drone car­ri­ers will be just as mo­bile, but longer-ranged and hit with more strik­ing power. And they’ll be much much cheaper as well. They’ll also be eas­ier to hide since they can pre­tend to be civilian trucks and eas­ier to de­ploy be­cause they aren’t as heavy or bulky and also the truck can po­ten­tially be stolen from the lo­cal pop­u­la­tion af­ter you ar­rive.

Fi­nally, this tech­ni­cally doesn’t count as a bat­tle bot be­cause it doesn’t fight, but… cur­rently ad­vanced mil­i­taries use a small por­tion of their sol­diers to shoot at the en­emy; most have some sort of job re­pairing and main­tain­ing ve­hi­cles, driv­ing trans­port ve­hi­cles, load­ing and un­load­ing things, etc. Many of those jobs could be au­to­mated away by self-driv­ing ve­hi­cles or boston-dy­nam­ics-style robots, mak­ing mil­i­tary lo­gis­tics faster and cheaper and free­ing up more troops for other tasks. This would be es­pe­cially use­ful if you need to main­tain a swarm of thou­sands of big drones.


Star­ships are ve­hi­cles like the SpaceX Star­ship. Ac­cess to space will be very im­por­tant in a fu­ture war. Be­ing able to view in real-time ev­ery inch of their ter­ri­tory, while they are stuck guess­ing what is hap­pen­ing in yours, is a hu­mon­gous ad­van­tage. If you can launch things into space more cheaply than they can, you can de­stroy their satel­lites and put up swarms of your own. Star­ships will plau­si­bly be at least one or­der of mag­ni­tude cheaper for de­liv­er­ing stuff into or­bit than legacy rock­ets, pos­si­bly two or more. (Caveat: In space, it is gen­er­ally eas­ier to de­stroy than to cre­ate. So even if one na­tion can put up 10x as many kilo­grams than an­other, the other na­tion might be able to ren­der space un­us­able, e.g. by cre­at­ing a sort of man-made Kessler Syn­drome. How­ever, with Star­ships it would be­come so cheap to put things into space that this difficulty might be over­come, e.g. by putting up swarms of ar­mored satel­lites that can last half an hour be­fore all be­ing de­stroyed, but dur­ing that time give you the valuable in­for­ma­tion you need. I’m not sure about this and would love to see an anal­y­sis of the effec­tive­ness of num­bers and/​or ar­mor vs. tons of tiny bits of metal pol­lut­ing LEO)

More ex­cit­ingly, Star­ship is sup­pos­edly go­ing to be ca­pa­ble of trans­port­ing cargo and hu­mans across large frac­tions of earth’s sur­face in mere min­utes. SpaceX is in talks with the US mil­i­tary about this; they claim that even­tu­ally their costs will be low enough that a Star­ship trip will cost about as much as a Globe­mas­ter (mil­i­tary cargo plane) trip. It can carry about as much cargo, but does so much faster and with­out the need of a run­way on the end. If this is true, it might make cargo planes ob­so­lete, or at least sub­stan­tially im­prove lo­gis­ti­cal ca­pa­bil­ities, es­pe­cially for na­tions who don’t have very pow­er­ful navies or air forces.

The most in­ter­est­ing (and also most spec­u­la­tive) po­ten­tial use is for de­liv­er­ing troops and other pay­loads into bat­tle. Be­cause the star­ship would prob­a­bly not be reusable af­ter get­ting shot at, and would not have ac­cess to re­fuel­ing fa­cil­ities in bat­tle any­way, this would be an ex­pen­sive strat­egy. How­ever, the benefits might be worth it—I’ll try to make the case for that be­low.

(You could use Star­ships that nor­mally serve as cargo ships, thus you still get to amor­tize the cost over sev­eral flights be­fore you throw them away in the at­tack. Or, you could cus­tom-build sin­gle-use Star­ships that use a two-stage de­sign to save all the ex­pen­sive bits for reuse.)

A SpaceX Star­ship can carry 100 tons. 50 tons of drones is plau­si­bly 50,000 drones of vary­ing sizes. That’s enough to par­a­lyze a small city; drones ev­ery­where, blar­ing on tinny speak­ers for ev­ery­one to lie on the ground with their hands on their heads. You’d have 50 tons of cargo space re­main­ing for troops and mis­cel­la­neous equip­ment.

Or, you might use 99 tons to carry ex­tra fuel, so that when you land you could im­me­di­ately take off again, per­haps af­ter hav­ing picked up or dropped off a few very im­por­tant pas­sen­gers. You wouldn’t be able to fly all the way home, but maybe you could fly some­where safe.

You could strike any­where in the en­emy coun­try, at any time, and they’d only have twenty min­utes of warn­ing at best. So, you could e.g. land at all ma­jor power sta­tions si­mul­ta­neously, or all ma­jor cities, or all ma­jor air­ports. (Seiz­ing air­ports is es­pe­cially im­por­tant be­cause you can use them to land re­in­force­ments; see the air­born in­va­sion of Crete) You could land wher­ever their forces are not, if you have good enough in­tel, which you prob­a­bly would since you have star­ships. Without ac­tu­ally at­tack­ing, but merely by threat­en­ing to, you could force them to keep much of their mil­i­tary at home, spread out over the coun­try, defend­ing var­i­ous im­por­tant in­stal­la­tions. This par­tic­u­lar ad­van­tage scales with the size of the coun­try you are fight­ing; larger na­tions have higher area-to-per­ime­ter ra­tios, so forc­ing them to defend their whole area rather than just their per­ime­ter is a big deal.

Be­ing able to strike any­where at any time on very short no­tice is more valuable than it sounds. It’s easy to un­der­es­ti­mate the im­por­tance of speed, sur­prise, and re­ac­tion time in mil­i­tary con­flict. For ex­am­ple, say a coun­try friendly to you is sur­prised by a coup. If you don’t act quickly, it will switch over to your en­emy’s camp. For­tu­nately, you can load and fuel a squadron of Star­ships in a few hours, and then it takes only twenty min­utes to ar­rive on scene. Or, maybe you are launch­ing a big at­tack at dawn. Si­mul­ta­neously with the at­tack you can as­sault the air­bases, anti-air fa­cil­ities, and trans­porta­tion hubs be­hind en­emy lines with drone swarms packed into star­ships. The drones could po­ten­tially be launched while the star­ship is still high in the sky, mak­ing mis­sion suc­cess still pos­si­ble even if the star­ship is shot down.

One prob­lem with at­tack star­ships is that they might be hard to dis­t­in­guish from nukes, and thus might trig­ger MAD. This prob­lem could po­ten­tially be solved by us­ing them against non-nu­clear en­e­mies, or by us­ing them one by one in­stead of all at once. (Launch one star­ship ev­ery fif­teen min­utes; that way if they are se­cretly nukes the en­emy will still have enough nukes left to mas­sively re­tal­i­ate; that way they won’t mas­sively re­tal­i­ate un­til they see whether they are nukes or not.)

3D printers

I am of the opinion that 3D print­ers will be a big deal in ten years or so. Cur­rently 3D print­ers are ex­pen­sive to buy, ex­pen­sive to op­er­ate, and the parts they build are crappy. So they are great for rapid pro­to­typ­ing, but not good for be­ing part of a fac­tory that makes a finished product. How­ever, the parts they build are rapidly get­ting bet­ter—in some ways they are bet­ter than any other man­u­fac­tur­ing method. Mean­while the cost of 3D print­ers is fal­ling pretty fast. Already I know of some prod­ucts (cars, rock­ets) that have some im­por­tant 3D-printed parts. As the price falls and qual­ity im­proves, more and more parts will be made with 3D print­ers.

Any­how, hav­ing lots of 3D print­ers ly­ing around is valuable mil­i­tar­ily, I think. Three rea­sons. First, mil­i­tary ve­hi­cles break down con­stantly and be­ing able to print re­place­ment parts is nice. Espe­cially since some ve­hi­cles will no longer be in pro­duc­tion so parts will be scarce, and other ve­hi­cles may be cap­tured from the en­emy and thus have no spare parts at all.

Se­cond, in­sur­gen­cies have sup­ply chain prob­lems. They need to buy weapons abroad and smug­gle them in some­how. Be­ing able to print what they need is a game-changer. Even ma­jor na­tions may have sup­ply chain prob­lems if their en­e­mies have de­stroyed your most im­por­tant fac­to­ries and sup­ply de­pots with long-range mis­siles or airstrikes; it sure would be nice to have a “vir­tual fac­tory” con­sist­ing of hun­dreds of thou­sands of 3D print­ers dis­tributed through­out your na­tion.

Third, a fu­ture war will in­volve rapidly chang­ing and evolv­ing tech­nolo­gies and tac­tics. World War Two was already like this; the weapons and tac­tics used at the end of the war were sub­stan­tially bet­ter than those used in the mid­dle, which were sub­stan­tially bet­ter than those used at the be­gin­ning. 3D print­ers speed up the de­sign/​pro­to­typ­ing pro­cess by, like, an or­der of mag­ni­tude or more. If you have enough of them, they can also speed up the mass-pro­duc­tion pro­cess as well; in­stead of tak­ing three months to build a fac­tory that pro­duces 1,000 of the new wid­gets per day, you can com­mand 10,000 of your na­tion’s 3D print­ers to im­me­di­ately pro­duce one wid­get per day each. Ex­am­ple: Your en­e­mies have coun­tered your drones by putting up nets ev­ery­where. Re­sponse: De­sign a net-cut­ter at­tach­ment for your drones, and patch their soft­ware so they know how to use it. Thanks to 3D print­ing you can be test­ing out 1,000 differ­ent de­signs within five hours of see­ing the en­emy nets. In a day you have set­tled on a de­sign that works, hav­ing col­lected enough data to train/​pro­gram your drones to use it. It’s a hacky solu­tion, sure, but that’s OK since you’ll be work­ing on a bet­ter solu­tion soon. By the next morn­ing you’ve printed a mil­lion net-clip­pers in var­i­ous fa­cil­ities around the coun­try and are already ship­ping them off to the front to be af­fixed to your drones. Another ex­am­ple: Your eggheads come up with a com­plex 3D pat­tern that works as an ad­ver­sar­ial ex­am­ple for the en­emy aim­bots, when viewed from a va­ri­ety of an­gles and light­ing con­di­tions. You print out a ba­jillion of them and dis­tribute them to your troops for the next as­sault, be­fore the en­emy re­al­izes the prob­lem and patches it. And when you find a new way to fool their aim­bots, your print­ers will be ready again.

Laser weapons

Their main ap­pli­ca­tion is for defense against in­com­ing drones and mis­siles (and bal­loons?). They are bulky and re­quire a lot of power, but they might still turn out to be bet­ter than the CIWS and other au­to­tur­ret sys­tems available. They could also po­ten­tially be used as bet­ter sniper weapons, since light trav­els much faster than bul­lets. They also can and already are be­ing used as “daz­zlers” to tem­porar­ily (or, if you are will­ing to vi­o­late treaties, per­ma­nently) blind hu­mans and cam­eras. How­ever I am not sure laser tech­nol­ogy will ever be good enough to out­com­pete more mun­dane al­ter­na­tives, and even if they do, they prob­a­bly won’t be that use­ful com­pared to con­ven­tional al­ter­na­tives. I might be wrong though.


Ex­oskele­tons are definitely a thing. More­over I am fairly con­fi­dent that given enough R&D, a ver­sion could be de­signed that in­cludes enough ar­mor to en­able the wearer to smash through doors and storm an en­emy-oc­cu­pied build­ing, in­vuln­er­a­ble to or­di­nary bul­lets. How­ever, by the time this is de­signed, it will be mostly ob­so­lete, as bat­tle bots can do the same stuff but bet­ter and cheaper and with­out risk­ing your sol­dier’s life. At least, so I pre­dict. Per­haps a lighter ex­oskele­ton will be built that goes over the sol­dier’s legs, and al­lows him to walk longer and carry heav­ier loads with­out get­ting tired. Even if this hap­pens, I doubt it will be a game-changer. Similarly, I ex­pect ex­oskele­tons might see some use for lo­gis­ti­cal pur­poses, helping troops load and un­load cargo, or arm and re­pair air­craft. But not a game-changer, es­pe­cially since robots might get good enough to do those tasks as well.