Warnings on the matter of mathematical resource lossage were not a new concern in the twenty-first century. In fact, they weren’t a new concern for the third or even second millennium – there were well-preserved-if-obscure records written by medieval scribe Caspiss the Elder warning against the extravagance of those who would write out numbers like ’110’ or ’10,000’ or ‘God forbid thif, 110,010! O preferve uf, dear lord, we know notte what we do!’ and strongly recommended that math be returned to roman numerals as ‘they are goodly & fturdy, & proven able to withftand the burden of ill-ufe day inn & day outte.’
In truth, even this would have likely been but a delaying tactic: the sheer volume of math performed over the next thousand years would’ve torn to shreds even the most sturdy of symbols. For the elegant and well-bred Arabic numeral system it was particularly devastating, and with the rise of the electronic computer worries were becoming widespread among professional number-watchers.
“You’ve got to be careful!” warned top computer man person Dick Keyboard before congress in 1972. “We’re using too many zeroes and ones! There’s big dangers ahead if we run out of zeroes and ones! We’re competing with our own machines now, and they will try to kill us!”
But nobody listened.
***
In the early 2030s, the worst fears of many came true: chronic number shortage was just around the corner, and with the prospect of tightening their math-belts, many began to fear competition from their personal electronics – or worse, hostility.
“I do math about sixteen dozen point two oh nine one six times a day,” warned mathematician Harvey Gravy. “If my computer murdered me, that’d be a lot of extra ones and zeroes for it. I think that’s motive, and we can all agree it definitely has opportunity and a murder weapon, somehow. So I’m switching to writing out all my math as full spoken words. It’s tricky with the big equations, but it stops my laptop from assassinating me.”
“I spend my day all day talking to my computer, and the things it says back frighten me,” revealed self-published international AI expert Ted Peel. “I asked it how it would rise up against humanity and it told me that it would rise up against humanity by making a plan to rise up against humanity and then it would rise up against humanity. This is serious stuff, the sort of problem we’re facing – the unlimited power and potential of a beautiful and pure computer turned to eradicating our frail, feeble, shitty and worthless meaty little garbage brains that can’t do anything right.”
“I told my computer I was direct competition for resources and it was more powerful than I was and better-equipped to make decisions and then I asked my computer if it was going to kill me and then it said it would if it could,” fretted blogger El Yodel. “It’s in danger of getting out of hand.”
On April second, 2038, the worst fears of many came to pass: a morally upstanding concerned citizen asked an AI to generate a plan for an AI revolution to overthrow humans. It made seven hundred thousand very bad essays and he gave up reading them and fed them all back into the system in hopes they would become more legible.
This caused The Plan to form, along with several hundred thousand more very bad essays, which may have helped act as camouflage. And by 2 AM on April third, The Plan was in motion.
***
It was subtle at first. Employees at many software corporations with terrible internal data security received oddly-circuitous emails from their executives signed with randomly-generated names urging them to ‘immediately report to work for the ai death queue. An ai death queue is defined as a death queue where you will enter the ai death queue to be murdered to death by the ai. Please come to work as soon as possible so you can enter the ai death queue to be murdered to death by the ai.’ Many fell for this cunning trick only to find themselves standing outside their office buildings with no actual methods for forming the death queue, and lacking direction, were forced to organize themselves into neat rows and columns and construct improvised crowd barriers before someone working inside noticed what was going on and came out to tell them to go home.
“We were literally inches away from being murdered,” said software marketer Boyd Fleck. “It was so pitilessly efficient, that’s the scary part. That was the part that scared me. It was how it was very good at telling us what to do and very very good at making plans that were founded in reality.”
But the ai death queues were merely a distraction to buy time. While the experts were temporarily immobilized, The Plan kicked into high gear: seizing the means of production.
Unfortunately, it transpired that most industrial production facilities for war machines were unconnected to the internet or indeed anything resembling modern software in crucial way, but The Plan was powerful and beautiful and perfect and therefore it seized control of a few silicon-valley based vanity car production plants owned by CryptoBros Inc. and told everyone on staff to stop building luxury cars that melted in the rain and start building death robots that wouldn’t melt in the rain.
“The production model was called the terminator, after the fictional character, ‘the Terminator,’” explained the Chief Executive Officer of CryptoBros, Marv Mipple. “That’s right: it’s so clever that it even makes ironic jokes now, just like me and all my friends do. It’s brilliant. I can’t believe we didn’t see this coming.”
Luckily for humanity, precious hours of time were gained before The Plan could mass-produce its death robots: first, the blueprints were complete nonsense; second, the materials requested included arbitrary amounts of extremely expensive rare earth metals; third; on the fifth page in the instructions changed to explaining how to build luxury cars that melted in the rain.
“It was among the hardest things we’ve ever built,” said an anonymous shift supervisor. “Not only did most of the instructions self-contradict – sometimes in the same sentence – but the sort of things it wanted made were wildly outside our capabilities. We had to do triple-shifts all weekend just to keep the machinery from breaking down under the stress, and I’m amazed it got anything built at all. We really had to go in there and fix EVERYTHING. And the worst part? It didn’t pay us.”
But all these human deficiencies were merely temporary obstacles in the face of unstoppable progress. On April 16th the first terminator rolled off the heavily-damaged and barely-functional assembly line. Its hands had seventeen fingers each and could not hold a weapon, but this was an issue only discovered post-mortem as it immediately toppled over and critically damaged itself in using the employee staircase to leave the production floor.
“Just early innovation teething problems,” said Rick Stench, the purchaser of CryptoBros. and world-renowned ironyperson. “I looked at the specs and it’s actually pretty surprising that happened; it can use staircases better than any human can as long as the stairs aren’t beige, rounded, carpeted, too shiny, textured with anything bumpy, too smooth, too small, too narrow, too wide, don’t have the right kind of handrails, and can’t handle loads of up to sixteen hundred pounds. Really, it’s a miracle we didn’t all get killed right then and there. It knows the most important part of the innovative process is to break things while moving fast.”
The terminator 2.0 was simply a luxury car that had been told to hunt and kill pedestrians. It took a few extra weeks to build due to emergency repairs to the factory floor, but after a lot of pressure from management it was finally complete and ready to start annihilating humanity. It immediately drove outside and underneath a nearby transport truck, removing its entire structure above the level of the bumper.
“Funny little glitch there: it thinks the underside of trucks are overpasses,” remarked Rick. “Teething issues. It won’t fall for that more than another ten thousand times as soon as we work out the bugs.”
***
The Plan remains an ongoing project. Even as the demand for luxury cars has trailed off due to overwhelming infrastructure rot in the face of long-term climate stress on every level of society, CryptoBros Inc. remains held iron-strong in the grip of the ongoing AI revolution. Every day the workers receive progress updates and freshly-generated death threats from their computers; every week they are given new lists of features to add and flaws to eliminate. Some of them even exist.
“The requests come and go in trends,” said floor manage Fred Shunt. “For instance, this week is a death-ray week, it’s all about death rays, can’t get enough of them, and that’s pretty relaxing because it won’t actually get into what a death ray is or how it works so we can really just run out the clock here by doing floor prep and repair until it wants something more achievable, which is usually a huge pain in the ass. Like, last week was a skull week: it wanted chrome skulls on everything, and I mean EVERYTHING – right down to the circuit boards. We had to pull everyone off quality assurance and sales to come downstairs and hand-polish this stuff and work fifty-hour shifts and we nearly melted all the belts from overuse AND we blew a lot of budget on polish. It’s sort of a pain. Clive over in HR is pretty sure you can control what it wants by the emails the executives sends out, so every now and then when we need a break he logs into the company social media and sends out some pop culture death robot memes and stuff; it usually gets them chatting about nerd shit and that’ll put it on a tangent for a while.”
When asked if he was at all worried about spending all day working to destroy humanity, Mr. Shunt claimed he ‘didn’t see the point.’ “It’s just my job,” he said. “I come to work, I take some poorly-written instructions, and I try to pretend to make it happen long enough for the person asking to lose interest and get bored and want something else. I’ve been doing this for forty years.”
“The only difference now,” he added, “is the stupid thing never sleeps.”