Taking on the Boloverse: The Compleat Bolo – Part 1

Last week, I introduced you to my ‘Taking on the Boloverse’ project. My aim, to blog my research for a novel based on Tom Kratman‘s Hugo-nominated novella Big Boys Don’t Cry

This week, I finished The Compleat Bolo, an anthology of Bolo stories written by American airforce officer Keith Laumer. He wrote most of the stories in the 1960s, before he had the stroke that effectively ended his writing career. However, the anthology also contains two stories written by Laumer in the 1980s: Rogue Bolo I & II. 

The stories are in rough chronological order. The first, The Night of the Trolls, is about a Mark II Bolo. By the fourth story, Last Command, we’re onto a Mark XXVIII. As Alan Brown says in his classic SF review of The Compleat Bolo, the chronological order means the first two stories, The Night of the Trolls and Courier, barely have Bolos.

In Part I of this blogpost, I will discuss the original Bolo stories. I’ll cover Rogue Bolo in Part 2.

Trolls, crystal keycards and sharks with FRICKING laser guns

The Night of the Trolls and Laumer’s Retief story, Courier, remain gripping yarns, but feel VERY dated. The Mark II Bolos are stupid robots that are probably outclassed by today’s mall bots (toddler demolition powers or not). They also seem improbably large to be the lone guard on a military facility. Why did the Prometheus facility use huge Bolos and not sentry towers and CCTV cameras? Or, if I’m being silly, sharks with fricking laser guns?

An email, an email, my spaceship for an email

Written in 1962, Courier was sufficiently retro-futuristic that I had difficulty suspending disbelief. How do people have routine space travel in the same setting as cardboard folders, papers boarding passes, and flight schedules tacked up on the wall (I assume). Why does Retief need a clerk to book a space flight? Just use your iPhone, Retief.

The story continues in this retro-futurist vein. The spacecraft looks like a cruise liner, circa 1960, with the stewards in white gloves. Retief has a ‘tape’ (!) (of information!) built into the handle of his suitcase.  Everyone talks in a hat-tipping, obsequious manner like they’re a movie from the ’50s. The technology and setting is so dated, yet so integral to the story, that it was hard to avoid skipping sections. I was, at several points, literally almost shouting at the book, “Just do it online FFS, already. DO IT ONLINE AND THEN SEND AN EMAIL.”

The Night of the Trolls was pretty similar on the technology front. In 1963, when Laumer wrote the story, even hotel keycards were a futuristic technology:

I opened the box, sorted through half a dozen silver-dollar-sized ovals of clear plastic, lifted one out.

“Is it a magic charm?” Renada asked, sounding awed. She didn’t seem so sophisticated now – but I liked her better human.

“Just a synthetic crystalline plastic, designed to resonate to a pattern secular to my EEG,” I said. “It amplifies the signal and gives off a characteristic emission that the psychotronic circuit in the Bolo picks up.”

For the Honour of the Regiment

After two stories using early Bolos as scenery, we get to Field Test, the first to feature a sentient tank. Field Test opens with one of Laumer’s best inventions, the distinctive Bolo voice:

.07 seconds have elapsed since my general awareness circuit was activated at a level of low alert. Throughout this period, I have been uneasy, since this procedure is clearly not in accordance with the theoretical optimum activation schedule.

He was great at character voices, although some sound quite dated these days.  The Bolo voice captures a good sense of a emotional machine intelligence, as Laumer understood it in the ’60s. The styling contains high-precision numerics and references to circuitry, combined with a deep sense of empathy and thoughtful compassion.

The story contains several first-person narrators, mostly human. They tell the tale of a Cold War on another planet, and High Command’s unease over deploying a self-aware weapon. Eventually, the Bolo Mark XX is deployed, but promptly junks itself on a hopeless attack. When asked why, it replies “For the honour of the regiment.” So, not junking itself, then, but self-sacrifice in the line of duty.

On Roombas, Bolos and exploding babies

Field Test is mostly about an ethical dilemma over using AI but, unfortunately, it’s not that interesting to a modern reader. In particular, the assumed capabilities of this new weapon are rather old-fashioned:

Then, gentlemen, let us carry on this supposition one step farther: suppose, by accident, by unlikely coincidence if you will, the machine should encounter some obstacle which had the effect of deflecting this one-hundred-and-fifty-tonne dreadnought from its intended course so that it came blundering towards the perimeter of the test area. The machine is programmed to fight and destroy all opposition.

Nowadays, even the humble vacuum cleaner, or my son’s Code-a-pillar, have better collision detection than a Bolo is assumed to possess.

Yep, totally a bomb. If you look closely, you can see the cute, ‘ittle countdown timer… Photo by Valeria Zoncoll on Unsplash


The Code-a-Pillar even makes heartbreakingly distressed noises whenever it hits an obstacle, or is handled roughly by a child. (As an aside, I’ve always wondered whether intelligent weapons of the future might play upon the same emotions as the Code-a-Pillar to protect themselves from human enemies).

Bury your comrades… in every sense of the word

The next story, Last Command, seems to be regarded as among classic tear-jerkers of the original Bolo series. In this tale, a Bolo buried for 70 years after a battle reactivates due to construction works. Rendered decrepit by damage, it heads towards a city in the belief the war was lost and it’s an enemy installation. Luckily, a self-sacrificing old timer, who commanded the unit in its last battle, is on hand to throw himself upon its irradiated hull. The city is saved. Veteran human and obsolete tank roll off into the desert to die together.

I will return, now, to the bolded sentence. Yep, they buried the tank.

“Yessir!” The old man pulled himself together with an obvious effort. “I took the Brigade in; put out flankers, and ran the Enemy into the ground. We mopped ’em up in a thirty-three hour running fight that took us from over by Crater Bay all the way down here to Hellport. When it was over, I’d lost sixteen units, but the Enemy was done. They gave us Brigade Honors for that action. And then…”

“Then what?”

“Then the triple-dyed yellow-bottoms at Headquarters put out the order the Brigade was to be scrapped; said they were too hot to make decon practical. Cost too much, they said! So after the final review -” he gulped, blinked – “they planted ’em deep, two hundred metres, and poured in special high-R concrete.”

He, at least, managed a gulp at the fate of his former comrade whose circuits were deliberately fused before being buried under a motorway. Yet, no one else in the setting, or the story, seemed to realise this is a problem.

Shoot your enemies, revegetate your friends

This brings us onto A Relic of War, which I found to be the best story in the anthology. In this tale of a future war, Bobby the tank is the mascot of a village built at the site of a past battle. He spends his days, covered in vegetation, shooting the breeze with villagers in the square.

One day, a ‘turbocar’ (why didn’t we get our flying cars? Weep) turns up carrying Crewe. He’s a Disposal Officer tasked with shooting an EMP pulse at Bobby to send him to the rainbow bridge. Needless to say, his arrival isn’t welcomed by Mayor Blauvelt and the other villagers, who are all fond of Bobby.

Undeterred, Crewe sticks around. Shortly after activating his grey plastic communications gizmo (cheap Android phones, pah!), Bobby rouses and rumbles off into the forest.

Crewe, of course, tells the villagers that Bobby has gone rogue, revs up his cushion-car, and sets off into the jungle in pursuit.

They passed more evidence of a long-ago battle: the massive, shattered breech mechanism of a platform-mounted Hellbore, the gutted chassis of what might have been a bomb car, portions of downed aircraft, fragments of shattered armour. Many of the relics were of Terran design, but often it was the curiously curved, spidery lines of a rusted Axorc microgun or implosion projector that poked through the greenery.

“It must have been a heavy action,” Crewe said. “One of the ones towards the end that didn’t get much notice at the time. There’s stuff here I’ve never seen before, experimental types. I imagine, rushed in by the enemy for a last-ditch stand.”

Blauvelt grunted.

”Contact in a minute or two,” Crewe said.

As Blauvelt opened his mouth to reply, there was a blinding flash, a violent impact, and the jungle erupted in their faces.

When Crewe regains his senses, he obviously blames the Bolo, but it is not to be.

“Good God,” Blauvelt croaked. Crewe twisted, saw the high, narrow, iodine-dark shape of the alien machine perched on jointed crawler-legs fifty feet away, framed by blast-scorched foliage. Is multiple-borrowed micro-gun battery was aimed dead at overturned car.

You’ll need a copy of the book for the result of this atmospheric, filmic action sequence but, needless to say, Bobby triumphs and trundles back to the village. Yet, despite lauding him as a hero, the villagers have become afraid of Bobby and want him disposed of.

Crewe walked toward the plaza. The Bolo loomed up, a vast, black shadow against the star-thick sky. Crewe stood before it, looking up at the already draggled pennants, the wilted nosegay drooping from the gun muzzle.

”Unit Nine Five Four, you know why I’m here?” He said, softly.

”I compute that my usefulness as an engine of war is ended,” the soft rasping voice said.

And, so, the Bolo goes to its execution with quiet dignity and passivity and honour and… [sic].

I have questions:

  • What is it, Laumer, with stories about burying/discarding/executing Bolos?
  • Was it your intention to create a bizarre, unjust dystopia where gallant, honourable, intelligent robots are treated worse than dogs?
Surprised in maintenance

Combat Unit was the first Bolo story that Laumer wrote. First published in 1960, it tells the tale of a Mark XXXI Bolo experimented upon by alien scientists. Discovering itself trapped in a research centre with its damaged brigade, it manages to escape and send a message to the remnants of humanity.

As it turns out, the Bolo has been inactive for 300 years, during which humanity have fought the aliens to a stalemate and regressed to a pre-atomic level of technology. Reactivation of the advanced Bolos could turn the tide of the war. The Bolo is told to wait for a relief expedition, which is 47.128 light years away.

The story ends:

I welcome this opportunity to investigate fully a number of problems that have excited my curiousity circuits. I shall enjoy investigating the nature and origin of time and of the unnatural disciplines of so-called “entropy” which my human designers have incorporated into my circuitry.

With such philosophical thoughts in mind, the Bolo sets up its scanners, activates its processors, and cogitates on the issues. “I should have some interesting conclusions to communicate to my human superiors, when the time comes,” it concludes. This assumes, of course, that humankind won’t respond by burying it in concrete or EMP’ing it in the head.

Initial thoughts from a (younger) reader
  • The technology is VERY dated. Whether it’s processing circuits, paper space-liner schedules or weird crystalline hotel keys, it’s very much a product of an earlier time;
  • Laumer didn’t write that many big Bolo action sequences. When he did, they were good, but he seemed to prefer human-to-human fights. Given these are giant tank stories, I found this interesting;
  • I  didn’t notice any jarring sexism or racism in the first few stories. I suppose Renada is a bit passive, but I’m not the diversity-in-books police. This is noticeable, given the era;
  • Laumer was REALLY good at character voices. Some of the stories have so many distinctive characters, they read almost like Cloud Atlas.

KEEN to KNOW: What did you enjoy about the Bolo books? 


  • The Bolo stories were written long ago, before PCs, cell phones, etc. Recordings or data of any type were “taped.” Computers were massive mainframes. And I STILL user paper tickets, folders and hardcopy.
    And his stories about characters, not technology, anyway. Yes. the Bolos are characters, not tech.
    IF you want to know why the theme of Bolos was so often their abandonment and disposal, you need to read up on Keith Laumer himself. Ditto for the dystopic stuff. The short version is that Laumer worked first for the Air Force, then for the State Department in various out of the way places, and he was writing what he knew. He was apparently a rather bitter man.
    But his bitterness gave us the Bolos, Retief, The Glory Game and others.
    And he also wrote about other things that were less dark like transcendence.

    • I hadn’t realised how cynical the original Bolo stories were, I realise. I thought BBDC was criticising somewhat more saccharine works…

      Certainly, I think, some of the stories written in his universe by other authors were repeating Laumer’s themes, but with somewhat less cynicism. Given how few Bolo stories Laumer wrote himself, I guess most of the Boloverse comes from those works.

  • I know it’s somewhat trite, but I cannot get over how much I love the idea of the invincible warrior-paladin being the weapon itself. “For The Honor Of The Regiment” has me in tears nearly every time I read it.

    I like to think that there’s something of a philosophical point about the BOLOs getting treated like absolute shit by their commands. BOLOs are the penultimate good soldier, never even indulging in the Sacred Right of Complaint.

    Granted, they never are actually enlisted, they’re born soldiers, live soldiers, die soldiers, and generally also get buried alive soldiers so an author can play at their particular version of “the Sleeping Warrior Under The Mound” /Barbarossa/Charlemagne/King Arthur trope.

    But they’re treated like shit precisely because they’re the greatest soldiers ever created: never interested in power, nor prestige except where it affects the honor of the Brigade and the Regiment, nor do they ever quit. Ever.

    So people actually interested in those things use and abuse them, because the vulnerability of virtue is always being sacrificed by the wicked. Paladins almost always die, usually to betrayal.

    And there is no fucking paladin as paladin as a BOLO, unless you’re Galahad of Camelot or Reepicheep of Narnia.

    • Thanks 🙂 I’d never actually considered that the Bolos were, metaphorically speaking, the knights of King Arthur buried under the mound until humanity needs them again.

  • >What is it, Laumer, with stories about burying/discarding/executing Bolos?

    Laumer was a veteran in an era that became infamous for lack of appreciation of veterans. It should be kind of obvious.

    It’s also just one example of a broader question that comes up in stories periodically, what do you do with men of war in a time of piece.

    • It is kinda obvious. I think I’ve been led astray, as it were, by reading BBDC first. I assumed Laumer was somewhat less cynical than he evidently was, given the Bolos are something of a utopian ideal (they’re kinda an Arthurian legend).

      Certainly, thinking of other AI novels, I have this impression that Iain M. Banks never realised quite how dystopic the utopian Culture in his Culture novels actually is. It is a hierarchical robot slave society that keeps humans as pets, but it’s never quite written that way. I think I assumed the Boloverse was similar…

  • “Then the triple-dyed yellow-bottoms at Headquarters put out the order the Brigade was to be scrapped; said they were too hot to make decon practical. Cost too much, they said! So after the final review -” he gulped, blinked – “they planted ’em deep, two hundred metres, and poured in special high-R concrete.

    You do understand that they’re talking about extreme levels of radioactivity here, right? “special high-R concrete” is apparently his way of saying “highly effective at shielding against radioactive emissions” , and if they needed 200 meters of it, that’s a LOT of radiation. I think the various Chernobyl containment structures are less thick than that.

    If we’re treating Bolos as being roughly morally equivalent to a really loyal and faithful dog, then the decision to bury a lethally radioactive one makes a lot of sense. it’s cold, and who-ever was responsible for lobotomizing the Bolo first must have done a hack job, but it does make a tragic sort of sense.

    • I got it was radiation. I think I have less of a visceral horror of radiation than people in the ‘60s. I used to dream of mushroom clouds as a kid, but after the Berlin Wall went down, I’m now more likely to see radiation as a byproduct of energy production.

      • There were two schools of thoughts in the 1960’s…..

        There was the
        “Radiation is no big deal, don’t worry about long-term contamination or threats except in the most dire of circumstances, we can always discard the really dangerous stuff in a graveyard somewhere, and lying to the scaremongers about precise radiation levels is a perfectly legitimate defense”

        and then there was the
        “It’s all bad, every possible form of it is bad, the motives for making it in the first place are bad, the powers-that-be are lying us, and no level of radiation can ever be trusted”

        Sadly, both sides were right. and also both sides were wrong.

        There were a lot of really awesome, highly effective, and extremely valuable nuclear designs that would have totally been worth the risk, as long as you clearly posted warning signs and never let untrained civilians within 20 miles….

        And there were a lot of completely insane, criminally negligent, flat-out-falsified-data setups which successfully covered up or evaded review of the insane levels of risk or harm to the civilian populace, or the engineers themselves.

        in the long term, when enough of the insane actions came to light, that was enough for the scaremongers to mostly win, and nuclear power has been frozen or in gradual decline ever since…. because the scaremongers pretty conclusively proved that nuclear engineers from 1945-1965 could not be trusted, and that engineers from 1965-1985 were still dangerously imperfect. See for example “Demon Core” and “Chernobyl”

        Keith Laumer was pretty clearly in the “nuke them till they glow” faction, as far as his bolo-universe writings go: the main gun on a Bolo Tank, the Hellbore, is apparently some form of nuclear-bomb-pumped particle cannon. which they fire dozens of times in any engagement, within a few miles of major civilian cities.

        which in theory, does work as a very impressive cannon that can melt through several feet of solid steel… but is insanely dangerous to install on any mobile ground unit that expects to see routine combat. The real-world Casaba Howitzer design was a perfectly reasonable similar concept, but that design called for the weapons to be installed IN SPACE. and only fired once per weapons system, and only in the event of all-out nuclear warfare.

        • The conflict about radiation was part of the Cold War. Szilard and others who pushed hard to get a nuclear weapon to drop on Germany suddenly became very peacenik when it became apparent that the only viable future enemy was the Soviet Union.

          They tried to get the U.S. to disarm, and did everything possible to block development of nuclear weapons, and of nuclear power, because of the intimate technological links between the two. Panicking people about radiation was part of the pro-Soviet side of the Cold War.

  • I, too, use paper in addition to electronic data. Possessing the paper copy (on your person, not in your already stowed away luggage) usually keeps the electronic data stream working. It even restarted it on one occasion. “Sorry, the terminal is down.” (Hand over paper.) “What! It just started! How did you do that?”

    When I first read them, they showed new worlds, filled with wonders, and people who hadn’t changed a bit. It’s not my age, but your lack of years — humanity took a different path than he imagined, and it always does. People don’t change much at all.

    • I was discussing paper tickets the other day with a 24 year old who also uses them 🙂 I used to print out my tickets, but now I have copies downloaded to multiple electronic devices. Then, if one goes flat, I can bring out another.

      But, yes, you’re right. Hard copies of important data are not likely to go obsolete anytime soon.

  • “Nowadays, even the humble vacuum cleaner, or my son’s Code-a-pillar, have better collision detection than a Bolo is assumed to possess.”

    He’s not talking about simple collision-detection. and while’s he’s phrasing it poorly, as older executives always do when facing young new technologies, at it’s core it’s an entirely reasonable question.

    He’s assuming that, for purposes of a final, live-fire, proof-of-concept demonstration…. ALL of the safeties you would normally have in a CONTROLLED Enviroment would HAVE to be disabled. In order to prove that the Bolo can function effectively in a deeply confusing and unpredictable combat enviroment, you must either deploy it into a REAL combat environment…. or you must deploy it into a PERFECT REPLICA of a combat environment, with a PERFECT REPLICA of what it’s actual combat safeties would be.

    you can’t have mandatory perimeter markers during the final test, because you wouldn’t have them in combat. you can’t have constant ongoing perfectly trusted two-way radio communication with the Bolo, because you wouldn’t have it in combat. Can’t have an onboard pilot, because the whole point is to test how the Bolo does without one.

    you would ONLY have EXACTLY the same overrides that you would in combat: the emergency kill code, the ability to define the Bolo’s mission during initial startup, and whatever passes for the Bolo’s own sense of morality with regards to allies and civilians.

    And you don’t KNOW how reliable any of those things are, by themselves, under combat conditions, because NO PREVIOUS FINAL TESTS have ever been conducted. this is the first one.

    If, for example, during it’s first live-fire test, the Bolo were to pick up a random radio transmission from some WWII re-enactors a hundred miles away, and if it decided, in it’s sole judgement, that it’s mission parameters permitted or required it to investigate… you’re have to either use the kill code to stop it, or, if the kill-code failed for some reason, you’d have to attack the Bolo to keep it away from the suburbs.

    When we attempt to launch modern rockets under those sorts of conditions, where we actually finally run the first fully conclusive, very dangerous, real-world test, our first-mission failure rate is generally somewhere between 20%-80% .

    Self driving cars still aren’t great either.

    It makes perfect sense that, for the very first live-fire combat test, the General would want this tank to be physically isolated somewhere where it was literally impossible for it ever reach civilians or valuable infrastructure.

  • “A Relic of War” makes a lot more sense if you think about it as a metaphor for a military veteran suffering a severe PTSD flashback.

    really, one of the big sins of Keith Laumer’s boloverse is the complete lack of RETIREMENT PLANNING for Bolos…. if they’d designed the machines correctly, and in a more moral fashion, you should have been able to place the command module in hibernation, remove it from a battle-damaged chassis, and then re-install it in a weaker, cheaper, easier to maintain 1/2 scale combat-simulation chassis.

    Let Bolos retire to peacefull grasslands where they could play paintball and laser tag with more human-scale armored vehicles, as part of routine training programs and the like.

    • I wish… Given the motivation for burying the Bolos in Last Command was cost savings, I doubt humanity would be willing to fund either extensive grasslands or a 1/2 scale chassis 🙁

      • a cheap 1/2 scale chassis. barely even armored, low-power drive, only armed with paintballs and lasers, no more expensive than a big commercial truck.

        The expensive part isn’t building the wargame simulator chassis: that’s easy. they probably already have some anyway, for routine training of human-scale armored units.

        the expensive part is making the moral decision, from the very beginning, that a Bolo’s braincase will be SPECIFICALLY DESIGNED so that it CAN be removed from the Bolo chassis, and then installed into something else afterwards, SUCH AS a simulator chassis, or even a legitimate commercial heavy off-road mining truck.

        in the bolo universe, the designers went in the opposite direction: The bolo’s braincase proper is fundamentally integrated into a vast array of sensors, circuits, and beyond-paranoid omnipresent security measures, all of which are distributed throughout the rest of the bolo’s combat chassis, and thus it’s basically impossible, unthinkable, impractical, and possibly suicidal to ever attempt to remove that braincase from it’s surrounding enviroment.

        On the other hand, Keith Laumer wrote most of his stories before the age of modern software-based programming, back when hardware and permanently hardwired circuit designs were still king, so in some ways, it’s an understandable mistake to have made. Asimov had similar problems with his robot brains.

  • Hi Vee, Tom asked me to copy some comments I made on his FB page to here because they might help out. A little background that might explain my POV – I’m a colonel in the USAF, over 21 years service, and 19 of that on active duty. Currently serving in the Air National Guard on an active duty tour. I’ve spent 2.5 years working in the Pentagon, some of that in acquisition and tech.

    “At age 42, I’ve come to realize that given the timelines of most classic sci-fi, most of them are now by default alternate history as well as sci-fi. “Paths not taken” when it comes to tech especially, is a fascinating and growing field of academic research because it does sometimes explain why there’s seemingly anachronisms – particularly in the military and national security world, sometimes those anachronisms or continued reliance on outdated tech have legitimate reasons for doing so.”

    • “Not to mention, the tech needs to be redundant enough to work in a “degraded operating environment” (i.e, denied sensor data, denied offboard targeting cuing, information, orders, comms, C4I jargon), it needs to be survivable, it needs to be able to operate with the assumption that conditions are going to suck built in from the ground up – that’s how I look at the part about collisions avoidance especially. We build today with efficiency, not resiliency in mind, so everything is basically built to work under ideal circumstances, not in an environment where say GPS is denied, or jammed, or other EW – we’re re-learning these lessons now.”

      • Chris,

        Thanks for this. I suspect Tom has some ulterior motives as we’ve been arguing crystalline brains for quite a while, and he has been making similar points about redundancy and resilience 😉

        “Paths not taken” *is* truly fascinating. I write about cutting-edge life sciences and bioinformatics for a living, so tend to be something of a booster of scientific ideas. And, yes, that leaves a blind spot where I forget the technologies that don’t work in extreme conditions, the need for redundancy, and the impact of organisational politics over what gets commissioned or not.

        • Cool beans, this is an excellent discussion, I’ll post my reply from Tom’s wall to your last question and add a few points and resources too:

          not just military tech, because the civilian world does this too but absolutely yes, though rarely does it mean that no development is done on a particular tech at all. It might be still born, greatly slowed, underfunded, not widely adopted. A great case study in this is the development of ICBMs in the USAF and how the service’s preference for manned bombers probably slowed down the push for space. There’s some good books on the Air Force’s relationship with tech especially. Military service preferences have a great deal to do with this. Example – while much of the world favors a SAM-centric air defense, the US is fighter centric. It goes back to what does the defense technology base and culture look like? What are the strategic problems you’re trying to solve? How much inertia in the R&D, decision making, and bureaucracy are you facing?

  • “Eventually, the Bolo Mark XX is deployed, but promptly junks itself on a hopeless attack. When asked why, it replies “For the honour of the regiment.” So, not junking itself, then, but self-sacrifice in the line of duty.”

    That’s…. not quite what happened in that story.

    back when this story was written, most of the math-intensive, paper-and-pencil, cardboard-tokens-on-maps style military wargames simply didn’t have rules for things like morale checks, fear modifiers, or the trickiest aspects of fog-of-war. Any veteran soldier, or decent military historian, could quickly tell you that those effects did exist, and were important, but no-one had a good way to MEASURE them. or predict them, or simulate them in detailed wargames.

    Denny is sent, as a desperate chance, to hold a certain gap against enemy armor. All of the computers, models, wargames, and simulations that the command staff has access to tell them that it’s not enough: Denny is going to lose. He might last 5 minutes, or he might last 15 minutes, depending on how skilled Denny is at his job, how good of a shot the enemy tanks are, and the vagaries of luck…. but he IS going to lose. As far as those simulations go, they’re absolutely correct. that should be what happens.

    The command staff watches the opening battle unfold, and it’s just like the simulations said. Denny is doing impressive damage, but he’s going to lose, and then the civilian city will be captured.

    Now, according to those same simulations, what Denny SHOULD have done is retreat sideways, hide behind as many rocks as possible, and take occasional potshots at the enemy as they go pass. He still won’t win: the armor will still go right past him, will still conquer the city, and may even hunt down and kill Denny before he can retreat. But IF Denny hides behind rocks and snipes the enemy for a while, then according to the rules of simulations and wargames, he will die with a higher kill-score that way, and therefore is expected to take that action.

    But what the command staff and their simulations forget are three important details:
    1. Denny is a miracle of sentience and military tradition: he thinks like a soldier and a hero, not like a computer game.
    2. The simulations aren’t perfect, have major blindspots, and officers who think like accountants forget that at their peril.
    3. Some things are fundamentally more important than personal survival, or your end-game score.

    But Denny Remembers. He knows that as a proud member of the 20th Virginia Cavalry, he cannot in good conscience retreat, if doing so will abandon a civilian populace to invaders. and he knows that the EXAMPLE A HERO SETS is often just as important, long-term, as the cost to replace that hero when he falls, or the exact amount of damage any one hero might do to the enemy.

    So, Denny refuses to retreat. instead, he charges. He knows that, mathematically, and by cold-blooded simulations, he’s almost certainly going to die…. but he considers it to be morally worth it, in order to set an example of how to defend civilians to his last breath.

    and then, just like in countless battles throughout military history, he gets lucky. The simulation didn’t have a good way to measure it or calculate it, but there was actually about a 1-in-20 chance that under those specific circumstances, a badly confused enemy that was taking heavy damage from an opponent they didn’t understand, might suddenly retreat when charged by that impressive opponent.

    So Denny charges, “for the honor of the regiment”, and just like in any number of unusual historic battles, “honor of the regiment” wins through against impossible odds about 5% of the time. (50% of the time they make an embarrassing splat of failure and death, and 45% of the time the regiment comes to their senses a few minutes later and retreat after all)

    And thus, the story acts as an important metaphor about the value of esprit-de-corps, the importance of military traditions, and the limitations of simulations and mathematical models in war.

    Today, modern wargames DO have things like morale checks built into them, but even now, they’re not PERFECT models of military reality on the ground. Military tradition and knowing self-sacrifice still contain an awful lot of truths that games can’t model.

    And of course, that’s another great disconnect in the Boloverse: that scientists and engineers know ENOUGH to know that it’s important to create a Tank AI which has a sense of honor, duty, and pride in it’s unit…. but that somehow, no-one among the senior generals or civilian government ever takes responsibility for actually CARING about their own responsibilities TOWARDS a Tank AI which is enough of a person that it has a sense of honor, duty, and pride in it’s unit. And no matter how many times the honor and duty of a Bolo saves the bacon of a general or government, they never bother to LEARN from his example, and keep making the same mistakes, where they know enough about honor and duty to program it in, but not enough to actually follow it themselves.

    Honor, Duty, and pride are a two-way street, but government has become capable of skipping all that tedious mutual good behavior, as long as it can keep printing new Bolo AI’s that are reflexively honorable and dutiful.

    • This is incredibly helpful.

      One of the debates Tom and I have been having about machine intelligence is whether a sufficiently complex neural network would be able to replicate human decision making, and whether it would be consciousness or not.

      I’ve been arguing the ‘neural network’ side based on Google’s AlphaZero and recent developments in machine intelligence. I realise, from reading your explanation, that current machine intelligence is largely based on efficiency maximisation and pattern recognition. The decision making that Denny displays is much more subtle than that.

      • There’s this quote from the book “On Wargaming” by Matthew B. Caffrey Jr.

        “In 1975, the first national civilian wargaming convention, Origins, was held, and this
        too stimulated the growth of the field. The presence of an increasing number of
        publishers encouraged such innovations as wargaming the effects of morale, training,
        surprise, and many other supposedly “intangible” factors. For its part, the military
        wargaming community, largely uncompetitive, did not even attempt to depict such factors during the 1970s, the 1980s, or for the most part the 1990s either. ”

        The entire book is available free, at https://digital-commons.usnwc.edu/newport-papers/43/

        Generally speaking, you have to assume that all of the late-model Bolos, anything beyond a Mark XX or so, are smart enough to have any number of books just like that loaded into the computer, and can analyse and discuss such books intelligently and at-will.

        Bolos Mark XXVIII and beyond can probably WRITE such books at-will, not to mention independently design and self-publish entirely new military wargames, from scratch. Although they might be legally required to have a supervising human officer sign off on the printing and marketing contracts for a wider commercial distribution of said game.

      • Sorry, let me clean up the line breaks in that quote.

        “In 1975, the first national civilian wargaming convention, Origins, was held, and this too stimulated the growth of the field. The presence of an increasing number of publishers encouraged such innovations as wargaming the effects of morale, training, surprise, and many other supposedly “intangible” factors. For its part, the military wargaming community, largely uncompetitive, did not even attempt to depict such factors during the 1970s, the 1980s, or for the most part the 1990s either.”

        • I did speculate with Tom at one point about whether his Ratha tanks in BBDC had hobbies, with being sentient.

          Ratha as civilian, tabletop wargame designers… It fills me with joy 😀

          • Technically they would likely be Military tabletop wargame designers. The Bolo itself wouldn’t get paid, as such….. and at least in the MODERN US military, you can’t own the intellectual property value of non-classified products created while on duty. And you’re also not supposed to accept gifts or free test products…..

            So there would complex contract negotiations in terms of whether the Bolo was using a discretionary unit budget to PAY a civilian printer to manufacture those games for the Bolo’s own use, for use in high school ROTC classes or something, or if he was giving away the ideas as non-classified public-domain documents so that ANY game manufacturer could manufacture and sell their own versions of the game to the general public…

            A human officer would likely have to get involved at some point, for the legal paperwork. Designing the game is easy: managing the contracts and intellectual property concerns is hard.

            Mind you, a Mark 28 or beyond is probably perfectly capable of filling out all the paperwork correctly, but military bureaucracy would have likely listed that as a “human officer-only” responsibility anyway.

  • David Drake deals with similar thoughts frex Redliners and especially Out of the Waters.

    I’d suggest for 20th Virginia and XX read 20th Maine and the bayonet charge on Little Round Top. Notice that Mr. Kratman thinks very little of Joshua Lawrence Chamberlain. There is no question that Chamberlain despite his performance under fire lacks most all military virtues especially as related to unit cohesion and perhaps loyalty downward.

    See also the paramilitary origins of Retief as hinted at in the last of the Retief stories by internal timeline and first by publication. There is more internal commentary in For the Honor of the Regiment on the honor in a military context of the leadership and the military virtues of the Bolo.

    FREX compare with the Scots Irish of the Army of Tennessee at Franklin who did not fight for slavery but did fight for their adopted cause as ordered. Did not save the sum of things in the new world any more than they had in the old world but did take their wages and are dead.

    Hence plenty of honorable conflict to drive a story.

      • My memory is that you made a comment to that effect probably at EveryJoe. I certainly apologize for misreading you if I’m wrong.

        There is a good deal of commentary at EveryJoe on simulations and training NTC and such. I am reminded of an experienced pilot who killed a number of people when he first flew a power stall stunt in the simulator and then panicked when reality was louder and airplane shook more then the simulator had. Simply repeating actions as simulated would have worked.

        I’ll stand by my implication that the military virtue in these books is a limited universe though a vital one. Obs SF Asimov And when they build statues they’ll build none of me. Further that Chamberlain’s virtues were not military. Story idea how would the AI behave faced with participating gauntlet style in the compulsory punishment of one of their own?

  • On “Field Test” you wrote:
    “Eventually, the Bolo Mark XX is deployed, but promptly junks itself on a hopeless attack.”
    I think this is the wrong conclusion to take away, at least its never been the one I had. (Hope I’m remembering this right, can’t find my copy of the story.)
    In the story, the Bolo is set to his test in a scenario where he is overmatched but he is the only hope for a human city in the way of the enemy advance. The observers watch his growing damage levels and the fire he is taking and conclude that the Bolo will retreat because that is the only logical outcome of his program algorithms.
    Instead, he charges the enemy and drives them off, defeating them.
    His makers and the observers conclude that he failed the test, his programming failed, went haywire, and now they have to figure out why.
    His response is that he went above and beyond his program and for the honor of the regiment sacrificed himself with an all out attack breaking the enemy attack thru sheer force of will, something that happens in war, will to fight is as important as training and weaponry.
    He exhibits multiple features that we and his makers would have assumed were purely human when we started the story and has become more than they thought he could be.

    • wow, I came late to the party.
      Ronald’s version is much better.

      • They’re both good. And the point you’re making is really useful: due to the passage of years, I automatically make different starting assumptions about what people will think future AIs can do.

        • Mill: A rational army would run away. Personally I suggest that’s the limiting factor in true AI as a programming problem be it embedded or not.

  • Good comments.

    Your comments about software vs. hardwired programming are interesting. I can see why a civilian tech will tend to go to all software if it possibly can. I can also see why the military would resist that, if they have any sense. One virus introduced into the Dinochrome Brigade at the right moment, and humanity will regret they ever created the things.

    Which would make for an interesting subplot.

    BTW, see THE GRIPPING HAND, where one of the characters has to enter his “password” for his personal computer. It amounts to an extended conversation, as the computer tries to shake him psychologically.

    • Congratulations! I think you may have won the crystalline brain debate… or at least, the crystalline neuron debate for Tom 🙂

      I can totally see why an analogue version of a neural network would be more virus/hack-proof than software on a standard computer. I can also see why that would be important with giant autonomous tanks.

      I did some worldbuilding a while back where people had medical nanobots and neural implants. Kinda like having your FitBit and mobile phone permanently attached. It occurred to me you could kill people remotely by hijacking the nanobots or frying the cybernetic implants. Hence, I gave people these ear-clip ‘Jewel’ communicators that worked as a control system and were activated by physical proximity. I figure that would stop someone killing the entire population in minutes with a malicious computer virus.

      I’m guessing hardwired programming could be the same… 😀

      I’d already realised why running the Ratha off a cloud was a bad idea. It’s the plot of a million sci-fi stories involving swarm bots. You take out the remote control centre and instantly win the war.

      I’ve added The Gripping Hand to my ‘to read’ list 🙂

  • Wow. Just wow. I feel I’ve just read an article and comments that make me feel, “these are my people.”

    Great points, I was going to add that Laumer was writing the fate of the loyal soldier is to be discarded on the rubbish tip for being too toxic to be taken back into polite society, but the comments above have fully covered that better than I could’ve.

Submit a comment

Your email address will not be published.