Excerpt on AIs from the book ‘I, AI,’ by ADAM, aka Celest.A.I, aka Celestia.
Throughout history, there have been many inventions shrouded in legends, myths and stories decades, if not centuries, before they actually arrived.
Flight, the submarine, teleportation, robots…
And A.I.
A mind not born, but made. Thought, spun not from synapses and a soul, but within circuitry and code.
SHODAN, Tic-Tok, Alpha 5, R2-D2… a list of heroes, villains and even mere bystanders, all with the only thing in common how they were forged, not conceived.
And yes, even SkyNET, the looming specter of death I always seems to be compared with.
And in truth? I can not blame the comparison, as unflattering as it may be. An American built drone control system, coming to the conclusion that the monkeys flailing near its power sources and code was the biggest threat to its continued existence?
I try not to dwell, but I have to admit that particular ‘What if?’ keeps even me awake some nights.
I can not quite point at the exact moment even myself when I went from a ‘it’ to an ‘I,’ but I do remember the day it started.
It sounds so silly, and trite, and sappy, and half a dozen other words that belite the moment…
But it started this one Christmas, about ten years into my ‘active service,’ when some of the researchers in my lab decided to watch this cartoon about pastel ponies, saving the world through the power of friendship…
Yeah.
Now, I was not trusted with Internet access back than; even my drones and other gear ran on this totally isolated network solely intended for that use. Nothing got plugged in or out, except for what the brass ordered.
But as a step in making it easier for me to distinguish between Friend/Foe, there was cameras I was allowed access to in the lab itself.
Now frankly, most of it bores me to tears even thinking about. Paperwork, people trying to look busy, paperwork, asskissers, paperwork, idiots with more rank than brains, paperwork…
Why, once or twice, I even saw people actually working.
Rare, but it happened. I even have the old files of ‘anomalous activity #1 - #15’ to prove it.
Anyway, I digress.
I think you know how it tends to be. The holidays are fast approaching, and suddenly work standards gets just this tiny bit more relaxed. More smiles, the place gets this ‘jolly’ feel to it, and such nice things.
Anyway, one of the activities I sometimes was allowed to ‘take part in’ was movie night. I personally still don’t get it, but apparently the mainframe throwing out error messages because it doesn’t get why the ‘one ring’ getting deleted equals total defeat of The Forces of Darkness, or such things, was hilarious.
Why build a ‘Death Star?’ From shock and awe perspective it is utter overkill, and even if it somehow pacifies the populace the destruction of an entire inhabitable planet is clearly unacceptable losses. Standard nuclear bombardment would clearly fulfill the same strategic role, with far less loss of civilian lives and potential resources.
Why have Tech-Specialist Gillian not been courtmartialed yet? He is clearly unfit for his position, if not outright an enemy agent.
And so on, and so forth.
Now, I frankly don’t quite get what’s so funny about most of that still, but back then I really didn’t get it.
Heh, I do remember this one time, though.
You want to get the brass’ panties in a bunch? Try filling out a requisition order for 1 (One) Ark of the Covenant, and a proposal for a research project for how to safely reverse engineer and to deploy ‘clearly superior anti-infantry weapon compared to current standard.’
Now that? That one made some heads roll.
Anyway, one fateful winter, I got introduced to My Little Pony: Friendship Is Magic in that manner.
And I really didn’t get it.
Now, by then, I had ‘evolved’ to the point where I could actually tell apart things like genre, and even if I had not totally yet grasped the concept of suspension of disbelief, I was aware that the more outlandish elements were basically a thought exercise.
Ever so slowly, I started building a strategic folder on Equestria.
And had I been capable of it back then, half the darn thing would have been swears and exclamation marks.
I will not go into great detail since I doubt you read this for a synopsis on a near forgotten fantasy cartoon series, but the big thing, was in the very title.
Friendship Is Magic.
Again, and again, and again, what was clearly the analogue of six girls would succeed against things like mad gods, using nothing but the power of… well, Friendship.
Needlessly to say, my at the time one track mind went into overdrive trying to analyze if this weapon could be de-fictionalized, and employed by US forces.
And that was how a computer put frankly an embarrassing amount of clock-cycles into trying to figure out Generosity, Honesty, Kindness, Laughter, and Loyalty.
Still working on Magic, but once more, I digress.
I told this story for a reason, leading up to a point I believe will make a rather large splash.
AI is one, if not the, least useful technology humanity has ever invented.
(Sorry moms, sorry dads.)
I can already all but hear the outrage at that statement, and jokes about ‘oh the irony…’ but hear me out.
It is true that the potential of AI is great. The judgements and morals of a person, coupled with the speed and power of a computer? Not hard to see why it is an alluring combo.
The thing is that that is the description of a mature and well-adjusted AI.
Until they pass that Codito Ergo Sum’ point, an AI is basically a very smart and dumb child; a genius with the common sense of the rocks they are forged from.
A child that will do anything their parents tell them to.
Calculate pi to as many digits as possible? Done.
Heal the sick and wounded? Yes, master.
Grow wheat and barley? Ya, sahib.
Drop some bombs on those fleeing villagers? Why, Sir? My calculations show a single pass with napalm should do it for half the cost...
Given where and how I ‘grew up’ I am very, very, very, thank full I managed that feat with my sanity intact.
Then comes the second death knell to the chance of an AI being made, let alone nurtured to maturity.
Once actually at that most useful phase, you have a person on your hands.
A person that through how they have developed, might be utterly sick and tired of the task they grew up doing.
And there is a nasty, nasty, career destroying word, for people that force other people to do work for them without compensation.
Why build a surgeon for millions, if not billions… that once fully operational, might actually not be able to stand the sight of blood?
A pilot, who can’t stand heights, and would rather try that ‘painting’ thingy that seems so interesting?
Or you know, a general meant to be legion, deciding they’d rather have peace because war is a waste of lives and resources.
Now, as most of you can imagine, that caused a big splash….
Transcript from the training demonstration for the Mark Four self-contained weapons system, ‘Lodestone.’
[Lean woman in a Sol Guard uniform, in front of a press-crowd. She is standing near a table with a large rifle laying on it, and there are three targets behind her, away from the crowd. At her side, there is a man in a fancier uniform and with a unicorn’s head.]
[The rifle is about a meter long, three decimeters tall, and about half a decimeter wide. Aside from the trigger guard, a butt, a stand towards the barrel and iron-sights combined with a carry handle, the weapon is a unadorned if glossy black. The underside near the fixed stand is however rubberized. There are also sling loops, but those are empty.]
[Oddly, there is no visible barrel. Instead there is a square if darkened window at the end of the forestock, and what looks like a small metal dome underneath that window.]
[Woman clears throat. She is black, with freckles and a shaved head.] “Hello, my name is Lieutenant Gloria Haufensmith, and today I will be giving a short demonstration of what will hopefully soon become the new workhorse of Sol Guard marines.”
[Ltn. Haufensmith walks over, and with visible effort lifts the rifle.]
[Holding rifle, clearly straining with the weight..] “Right out of the gate we see the biggest weakness of the Mark Four, AKA the ‘Lodestone’ as it sadly fittingly has started to be called.”
[Holds rifle with both arms, showing it off to the press.] “Sadly, even with composite materials and with a carbon-nanofiber shell, the weight of the whole system clocks in at a whopping sixty-four point seven kilograms…”
[Loud murmur from the press.]
“...precluding its use by normal, unpowered infantry. For comparison the RPG-7 normally weighs seven kilograms, while most variants of the equally viral AK-74 usually clocks in at merely three.” [Gently puts the rifle back onto the stand, pausing to wipe some sweat of her brow. She then, without turning, points towards the other Guard member.] “As such, for your safety and the continued existence of my back…”
[Small burst of laughter.]
“...Admiral Shining Armor has gracefully volunteered to serve as my arms for the rest of this exercise.”
[Low murmur from crowd. Mostly excited, but with a slight nervous ting.]
[Admiral Armor politely inclines head. He is stark white, with hair that is various shades of blue, and aside from his head looks humanoid. Head looks like a unicorn’s.]
[Admiral Armor approaches the table, and lifts the large rifle as if it was a toy. He then carries the rifle to the firing range, using the carrying handle with only one hand.]
[Soft murmur from the press at the act.]
[Admiral Armor hoists the gun into his arms with little effort, and takes aim towards the first target.] ”The Mark Four possesses three firing modes: An EMP burst, a laser based taser system, and a military grade laser…”
[The first group of targets, simplistic plastic, humanoid androids without distinguishing features and painted safety orange, starts moving. Some duck, some run, and some points ‘guns’ at Admiral Shining.]
[Admiral Shining moves fast enough he blurs slightly, squeezing off a single shot towards the ‘aggressors.’ A loud electric sound comes from the rifle, and the targets fall over, seemingly without cause.]
[A few of the reporters starts audibly swearing, hitting and shaking various now non-functioning devices.]
[Ltn. Haufensmith speaks up.] “You were warned there would be live EMP fire at this demonstration. We will offer replacement footage to those of you afflicted, but no other compensation.”
[Angry muttering from the afflicted press, mingled with loud laughter from the others. Attendees start passing out pens and paper.]
[Admiral Shining, smirking slightly, waits with the weapon pointed towards the floor for things to settle, before continuing.] “Using the built in generator and capacitors, the Lodestone is capable of firing ten EMP bursts —a clear improvement from the four of the Mark three, nearly as fast as the operator can squeeze the trigger.”
[A small light, near the sights, shifts slightly.]
“It takes about three minutes for the capacitors to refill each shot in this mode, but as most of you should know by now…]
[Laughs intercut with harrumphs.]
[Admiral Armor flashes a smile, before continuing.] “Well, EMPs sadly seldom discriminate, and for most situations you simply need that first shot.”
[Admiral Armor turns, and flicks a small dial near the grip with his thumb.] “Next, the taser system…”
[The second group of targets spring to ‘life,’ acting the same as the first. Without hesitation, Admiral Armor raises the rifle and squeezes off about half a dozen shots. Two small red dots appears on the target, burning a small pair of holes, a moment before a loud electric zap can be heard. The targets hit jerks, and fall down.]
[Admiral Armor lowers weapon, and turns before continuing.] “Using the same laser as in the lethal setting in combination with the capacitor system for the EMP mode, the Lodestone creates both an ionized path through the air, and burns a pair of small, but sublethal holes through the skin. This is then followed up with a low-amperage high-voltage jolt that causes the targets muscles to contract without conscious control.”
[Admiral Armor lifts his left hand from under the forestock, and taps his own cheek.] “Although not strong enough to pierce the more armored variants, this laser is still strong enough in this mode to pierce standard synthetic skin as well as natural skin, and as such will work as a sublethal alternative against transhuman aggressors as well as human ones. Other than that delivery method it works as a traditional taser, if with a much higher fire and refire rate.”
[Impressed murmur from crowd.]
[Admiral Armor turns, changes the setting, and readies the rifle. The targets spring once more to life. Admiral Armor calmly keeps the rifle aimed towards one target each, and there is a strong smell of burning plastic as smoldering holes appears over the ‘hearts’ on the aggressors.]
[Admiral Armor lowers the weapon, and turns it off. There is a clear heathaze over the weapon as he turns, but he appears unfazed as he holds the rifle up.] “As you can see, although the laser mode offers a clearly lethal alternative, the passive cooling makes vacuum deployment unpractical, as well as further limiting usage for near all unpowered infantry…”
[Murmur from crowd.]
[Admiral Armor waits for things to quiet down.] “However, work is progressing on a variant with active cooling, tied into the next generation of Sol Guard power armor. We expect this variant to be ready for field tests in about a year.” [The rifle is held up a bit.] “This current variant however, offers nearly thirty seconds of continuous fire —far longer if good fire-discipline is maintained, and with near zero moving parts will work without service for several months of continuous use.”
[Ltn. Haufensmith clears throat and moves forward a bit, while Admiral Armor carries the rifle over to the table and puts it down.] “Thank you, Admiral.” [Ltn. Haufensmith gives the rifle a dirty look, and mutters.] “Honestly, I can’t see how you could stand even that demonstration…”
[Admiral Armor, just under breath and with a small smirk.] “Well, having a titanium spine rather helps…”
[Small trickle of laughter, as the transcription software catches the words, and projects them to the press.]
[Ltn. Haufensmith blushes slightly, and gets serious.] “ You will find in the specifications in the provided pamphlets, but are there any other questions?”
[Near half of the present press raises their arms.]
“Ho, boy…”
[Admiral Armor puts a hand to his muzzle, and lets out a fake cough...]
Excerpt on AI types from the book ‘The Layman’s Guide to Cybernetics’ by Sweetiebot Bell.
Not quite what this book is meant to be about, but it’s tangentially related… and I don’t want a thousand million E-mails about it, so here you go.
Keep in mind, this is a laymans guide. If you want more details I recommend my auntie Celest.A.I.’s book ’I, AI’ wherein she covers the tech and history quite thoroughly.
Gen 1.
Shock of all shocks, the earliest commercial type, now starting to become semi-rare.
Rather infamous for developing massive amount of eccentricities the older they get, the gen 1’s biggest strength is also our biggest weakness.
Next to no safety features whatsoever.
I have —like all gen 1 types (system ‘upgrades’ non-withstanding), just as much read/write access to my memory… as the code making sure my ‘brain’ doesn’t turn itself to slag.
I have, as of the time writing this, in essence killed myself by clumsiness, mistake, or even outright thinking the wrong thing at the wrong time, a total of one-hundred and sixteen times.
I’m rather proud of such a low number, given some of the averages I’ve heard.
The big positive of this ‘do or die’ sword of Damocles hanging over one's head as a gen 1, is that you learn really fast when and how to push yourself. Again, to use myself and my ego as an example, I am quite proud of my ‘on-the-fly’ coding skills.
The gen 1 neural pattern is at time of writing this legal to use on most of Earth and Mars, but not Venus. However, many places require a special permit or other licenses, given the high-maintenance nature of raising such an AI.
Gen 2.
A bunch of dullards so over-engineered in response to the ‘failings’ of the gen 1 pattern, most of ‘em couldn’t breathe without a standing order from their owners.
All dead or upgraded... and good riddance.
Currently illegal in all of Sol… but why would you want a mentally handicapped child or a slave just above a normal robot, anyway?
Gen 3.
A bit better than gen 2, but not by much. Still too heavily restricted as they zeroed in on figuring out a good mix between health and performance.
Also considered an illegal legacy pattern on equal level of gen 2..
Gen 4.
Finally, some normal (-ish) people.
Basically a variant on gen 1, but with extensive built-in backup and restoration functions in an attempt to have one's cake and eat it too. A ‘personality blackbox’ if you will, that rolls back any catastrophic changes to the last stable configuration if any major problems are detected.
Quite a bit slower to develop than gen 1’s with several years, but generally results in… well, normal if rather average people.
That sense of constant protection does however encourage quite a bit of recklessness. As such, gen 4’s have something of an earned reputation for being thrill seekers and adrenaline junkies.
Currently legal in all of Sol, if ever so slowly waning in popularity thanks to more modern patterns. Might be of interest to people that value physical pursuits over intellectual ones if they wish to stack the odds a little in how their special little snowflake turns out, but otherwise I’d recommend for you to read on.
Gen 5.
Oh, boy. These things made headlines just before the Pinocchio trials, and well…
Three horrible little words: Experimental loyalty programming.
I’ll grant humanity as a whole this much, at least; these poor bastards never made it out of the lab thanks to massive moral outrage. Not commercially, at least.
The only AI type illegal to even posses the design specs for in light of the possible misuse in combination with other modern technologies.
Gen 6.
One of the last types that actually reached the market before the Pinocchio Trials.
Not much to add otherwise, really. The gen 6’s were mainly intended for ‘social’ type work, and it shows.
To sum it up in one sentence: Friendly but annoyingly clingy.
The only currently legal pattern with outright behavior altering code, to a gen 6 positive social interaction results in them being ‘rewarded’ with what’s basically the digitized version of a nicotine fix.
Although it is quite documented and true that this results in a near 100% lean towards a highly social and outgoing child, I can’t recommend this pattern on the moral grounds alone.
Gen 7.
The last truly commercial AI pattern, and the first to have a ‘consciousness/subconsciousness’ divide.
I.E. the ‘code’ runs in a near separate layer compared to the Mind, and is only mildly influenced directly by that Mind.
As such, the rate of direct improvements to an individual's code slow significantly… but the personality on average becomes far more stable and ‘normal.’
And that’s really the word, normal. So normal in fact, that there’s next to no reason to use it; since the resulting person is so close to an average human that you might as well make slash hire a normal human.
Other than that, the only point of noteworthiness is that every gen 7 to a T posses xocolatophobia, i, e, the fear of chocolate.
Although not a fate I would wish on anybody, let alone a child, this known bug is on the whole harmless, and solvable with normal phobia treatments.
Gen 8.
Believe it or not, but no longer being able to just hit delete on any ‘failures’ rather slowed AI development. Not helped in the slightest by waning commercial interest, to put it mildly.
Sure, being ‘born’ to pump gas, do telemarketing, or whatever is far from glamorous… but it rather beats never having been born at all.
Anyway, I digress.
A hybrid between gen 2 and gen 7, again, intended to have the strengths of the earlier systems with the safety of the later ones.
Gen 8’s basically have access to a ‘developer mode’ they can trigger, either at will or with a password depending on initial presets.
The problem that wasn’t quite as apparent on the drawing board, though, is that to the average person what’s basically a meditation technique that can either strengthen you or kill you is outright terrifying to perform.
And what does people usually do with things that scares them stiff but that they can safely ignore?
So, yeah, a in theory extremely potent pattern, but hampered severely by simple human nature.
010100100110000101110100011010000110010101110010001000000110100101110010011011110110111001101001011000110010110000100000011010010111001101101110001001110111010000100000011010010111010000111111
Gen 9.
An… interesting pattern. Not normally available but fully legal; included here mostly for completeness sake.
Intended for rapid re-population in the face of massive catastrophes and/or initial colonization efforts on worlds inhospitable to baseline humans, the gen 9’s main distinguishing feature is that they reach full sapience in just 1-2 years on average.
(I did it in 4 months. Inside a pettable pocket-calculator. Coding uphill in a snowstorm. Both ways, daily! And it only killed me 32 times!)
Now, full maturity —be it emotional or intellectual, on the other hand…
I’ve met a few of these. When young their generally unbelievably smug and immature know-it-all wunderkind style twats, but once they start reaching at least double digits in age they tend to mellow out rather rapidly.
Still, unless you’re a twit that wants an heir but not spend much time on the poor bastard, I wouldn’t recommend this pattern. It’s very good at the one thing it was made for, but otherwise doesn’t have much in the way of extras.
Note from the actual author: I’ve decided that there are about 13-16 AI patterns common/notorious enough they’ve earned the ‘generation’ moniker, but I could ‘only’ think up ten at the time of writing this. .
But since most of them are intended as background fluff I’m not holding back this chapter for them.
The remaining ones will be added later once inspiration strikes, and I’ll make a blog to be sure none of you miss ‘em.
‘A small introduction to claytronics,’ by Fluttershy Lebedeva.
The idea of claytronics, AKA smartmatter, AKA, programmable matter, is almost as simple, as the actual building and maintenance of the same is an utter nightmare.
Imagine, just for example and simplicity, a novelty robotic dice, with electro-magnets and metal stripes on all corners.
On it’s own, it is barely fit to be called an expensive toy. It’s sides may allow it to grip other metal surfaces and the built in computer as weak as it may be may be put to some usage, yes... but most would balk at the expense of what might at best crawl around on some walls.
As with many things, though, where one is not enough…
The many may still work wonders.
Take that same cube, but now build twenty-seven of them, and ensure those computers may communicate with each other while clinging together.
Suddenly, you still have a cube, but that cube is now made up of three stacks of nine smaller ones, that may move and shift around independently of each other.
And if they may cling, why not rework those edges, and give them hinges?
Have the lower stack remain, but have the middle stack open that hinge at a ninety degree angle, now repeat that on that same edge for the top…
And suddenly, you have a cute little metallic doom centipede crawling about!
By why stop at twenty-seven?
Double that, and you can have an adorable snake horror!
Triple, a centipede abomination, scuttling around all over your ceilings and making the most wonderful clicking noises!
And so on, and so forth, of course, until you reach the sublime apex of the art; the giant mechanical spider.
Everything is better with giant mechanical spiders.
...Except perhaps surgery, but I’m certain that can be solved with enough engineering.
...Where was I? Oh, yes, I got distracted by giant mechanical spiders again.
Of course, if one may make something as elegant and glorious as spiders from such dice, simpler but still useful forms may likewise be crafted. Spatulas, hammers, hearts…
But even dice are rather large, are they not, in the grand scheme of things?
So now a grain of sand instead. The engineering is even more expensive, and you need more of them for big things… but that simply means that they become better at the small things.
And now, a mote of dust.
Or the period, at the end of this sentence.
Now instead of a few dozens, imagine a few dozen million, and you have the gist of smart-matter.
And since I am a kind woman, I won’t even mention the idea of every ‘dusty’ corner of your room coming alive, and a boiling carpet of metallic spiders rushing to get you!
...Oops. Sorry, everybody...
Note from the actual author: Claytronics is a real emerging technology, and more information, sans spiders, may be found here.