Ex Nihilo

A story on endings and beginnings

Hello everyone, I’ve taken a break from posting on Substack since I wanted to focus on writing a longer-form piece, which ended up being today’s 15-minute short story.

Hope you enjoy.

---

Photo by Lukas Robertson on Unsplash

Humanity’s end began in 1807 in some inconspicuous cafe in the quiet Parisian suburb of Montmartre.

The great mathematician Carl Friedrich Gauss was walking through the narrow, cobblestone roads of Montmartre with his wife Johanna. Prior to their trip to France, Gauss had been working a brutal schedule, both teaching at the University and doing his research late into the night.

Johanna had thought that a trip to Paris would be a great way for them to spend some more quality time together. The vacation had been going quite well until something inside a cafe caught Gauss’ attention.

“My goodness, is that Legendre?” said Gauss. His face, already reddened from scaling the hills of Montmartre, was turning a shade redder, something that caught the attention of his wife Johanna.

“It’s been two years Carl, we really must get on with it,” Johanna said with a forced smile.

She couldn’t help but think of what awful luck it was to run into Legendre on this otherwise perfect vacation. In a last-ditch effort to distract Gauss, she exclaimed, “The cathedral is right up the hill. We’re almost there!”

Gauss scoffed. “It’s not just that he stole my idea, Johanna. It’s how he did it. The man practically gloated.”

Gauss, you see, was upset about Legendre claiming the discovery of the method of least squares. To explain the concept simply, the method of least squares would describe the slope of a line that cuts directly through the center of a scatterplot.

Gauss had known of this computational method since 1795 and had used it in several papers he wrote on planetary motion. Gauss, however, didn’t go out of his way to highlight his usage of this apparently groundbreaking algorithm.

But last year, Gauss began to hear whispers about a French mathematician Adrien-Marie Legendre who had developed an interesting tool for optimization, a tool that appeared eerily familiar to Gauss.

As Gauss entered that cafe in Montmartre, he remembered that feeling of familiarity upon reading Legendre’s paper and how that feeling turned into rage, not unlike the anger he was feeling at that very moment. Gauss huffed his way to the corner of the cafe where Legendre was seated.

“Legendre, you might remember me, we met many years ago at that conference in Munich.”

Legendre looked up from his newspaper, smiled, and surprised Gauss with a hug normally reserved for longtime friends.

“Ah Gauss, yes it has been quite a while. And this must be your lovely wife. Please sit down, both of you. Welcome to my home country. I know it must have been quite the trip from Germany.”

Born in 1752, Legendre was over two decades older than Gauss and a more well-recognized mathematician at the time. Unlike Gauss who had been born to a poor, working-class family, Legendre had been born into the aristocracy.

Private schooling, the university, and aristocratic connections were just a few of the advantages that Legendre had readily available to him. The world seemed to bend itself to Legendre. Unfortunately for Gauss, he was yet another person on the wrong side of Legendre’s relentlessly good fortune.

“And tell me Gauss, what brings you to this lovely cafe in Paris today?” Legendre asked as the couple sat down.

Gauss took a deep breath, steadying himself. “That paper you wrote on what you called the least squares method? That is a computational method I have been using for quite some time and -”

“Hold on there, friend,” Legendre said with his hands raised, “I came up with this method myself while collaborating with other mathematicians in Collège Mazarin. I can give you their addresses. Write to them, and they will attest to this.”

“I can show you papers published five years ago that demonstrate me using this method. I can’t help but wonder if my papers precipitated this wonderful discussion you had at Collège Mazarin.”

Legendre chuckled. “Very good Gauss, very good. It appears we are at an impasse. I will say that I didn’t read your papers, and you will insist that I did. But, my friend, there is one matter that we can agree upon.”

“The matter being?”

“That the whole world recognizes me as the creator of this computational method.”

Gauss looked down at the floor and ground his teeth. The old man was right, he thought.

Legendre continued. “Your problem is that you didn’t state - or recognize - the sheer potential for this computational method. You can use the method of least squares to make a prediction on any new input value.”

“I used the method in a variety of planetary motion problems, which is an extensive-”

“And that’s all it was to you. A way to solve astronomy problems,” Legendre leaned in. “But the potential for this method is infinitely more than that. You see, all we perceive in our lives is data. And we act on that data. In a way, the method of least squares is a very rough attempt to approximate some aspect of our consciousness.”

At this, Gauss couldn’t contain himself and laughed.

Johanna, meanwhile, was intrigued and asked, “So you think consciousness is just... a series of math problems to solve?”

Legendre smiled. “Precisely. When you walk to the market, you’re doing math to take the shortest path there. When you’re deciding what to do on a free Sunday afternoon, you’re doing utilitarian calculus to figure out what brings you the most joy. The key insight to least squares is that in our lives, we are either minimizing or maximizing. That is all.”

Gauss shook his head vehemently. “Now, you’re extending yourself Legendre. God, love, revenge, jealousy... There are things you can’t quantify. There are questions that can’t be framed mathematically. You must know this Legendre.”

Legendre was about to respond, but Johanna interjected, “You say this with such certainty honey, but who knows what’ll come in one hundred years? Two hundred? Or four thousand?”

Johanna shook her head. “It’s just hard to know for sure.”

Legendre nodded. “Your wife is wiser than the two of us.”

“I suppose we can all agree on something,” Johanna said glaring at Gauss.

Gauss sighed and reluctantly kept his mouth shut. It was the least he could do after initiating this unpleasant detour. The trio parted ways a few minutes later, once it was clear that Gauss no longer wanted anything to do with Legendre.

It was on this day that consciousness was first postulated as a series of optimization problems.

---

Since that fateful day at Montmartre, humanity continued to press forward in the computational space. The method of least squares gave way to more complex optimization techniques that eventually became incomprehensible to even the mathematicians creating these algorithms. However, with incomprehension, came computing power.

In the 21st century, the intelligence of these algorithms was put to the test in a series of Go matches between Google’s deep learning algorithm AlphaGo and Lee Sedol, a world Go champion. Go, an ancient Chinese board game, was long seen as a game that was more artful than analytical, given its sheer complexity. Yet, AlphaGo beat Lee Sedol in four out of their five matches, sending the world champion into early retirement.

AlphaGo had proven Legendre right to some extent: no matter the complexity, a problem could be solved as long as it could be framed as something that could be minimized or maximized.

With this insight, humanity began to reframe long-standing tasks into math problems that could be solved by algorithms, which soon began to be referred to as AI. But as AI crept its way into more sectors of society, economic inequality increased dramatically. In the same way that factory owners reaped the rewards of the Industrial Revolution, those who had access to AI-based tools hoarded the productivity gains of the AI Revolution.

Governments, keen on keeping a lid on growing inequality, passed laws forcing companies and universities to ban the use of AI in merit-based processes. These private organizations, however, had trouble detecting the use of AI. Well to do students continued to use AI-based writing tools for their essays. Wealthier job applicants still used these tools for their cover letters.

It was in this context that two young entrepreneurs created an AI-auditing software with the goal of creating a level playing field for humanity.

This AI-auditing software would prove to be a turning point in the development of AI.

---

“And we’re live at CNBC with Josh Phillips, the co-founder of Audit. Welcome, Josh.”

“Hey Alex, it’s great to be here.”

“So let’s go ahead and get started. Your company came out of the blue a few years ago when politicians from both sides of the aisle began praising your company. How did that happen?”

“Equality of opportunity is maybe the one thing that all Americans want. And with the high cost of AI-assisted writing tools, equality of opportunity was becoming a myth. And that’s why we started Audit… so we could identify AI-written works and create a level playing field.”

“I had never heard of you guys until last year when my daughter’s college essay needed to be accredited by Audit. And all of a sudden, you guys are used everywhere and have ballooned to a $20B valuation.”

“That’s right Alex. We want Audit to be embedded into any process that requires content to be written by humans. College essays, take-home interview assignments, you name it.”

“But you do have some detractors. For example, Magnus Carlsen, one of the greatest chess players in the world, claims that advancements in chess theory resulted directly from analyzing moves made by chess-playing AI. Do you think that the writing world is being held back by your company’s mission to promote human-only content?”

“I think the writing world is thankful that Audit is here to ensure an even playing field. Not everyone has access to these AI-assistant tools, Alex.”

“Fair enough. Josh, thank you again for coming on.”

“Glad to be here.”

Josh unclipped his mic and out of the corner of his eye noticed Delilah, waiting by the cameraman.

Josh and Delilah founded Audit together two years ago, and that was about all that they had in common. Josh, being the extrovert, was the CEO and the external face of Audit. Delilah, more soft-spoken by nature, took on the CTO role at the company.

This partnership worked well, to say the least. After all, they both grew the company to a $20B valuation in a fairly short amount of time. But over the past six months, tensions were building and as Josh walked towards Delilah, he knew what she would want to discuss.

“We need to talk,” Delilah said quietly.

“About what?” Josh said sarcastically.

“Funny. We’ve run some numbers to quantify my team’s concerns.”

Josh sighed. “Let’s go for a walk.”

The two went outside the recording studio and walked through San Francisco’s Chinatown in silence. It was a Saturday morning and the streets were bustling with tourists and locals alike.

“Listen,” Josh started. “I’ll say what I told you before. We can hire more auditors with Bachelor’s degrees. We can even expand the pool of Ph.D. auditors we have. As long as we improve our data-labeling process, our detection rate will improve. That’s what you’ve always told me, right?”

“Josh, our detection rate has been nosediving for the past 6 months now. Sure, we can hire more experienced auditors and stem the bleeding for 3 months. But what about after that?”

“We’ve gotten this far, using a combination of humans and AI to detect AI-assisted writing -”

“In two years, we’re not going to be able to tell the difference between human and AI written content. And that’s our whole business model.” Delilah sighed. “Let me talk at you for a bit, Josh.”

“Alright, let’s hear it.”

“In Romeo and Juliet, there’s that character Mercutio, Romeo’s fun and witty friend. Remember him?”

“Sure, ya.”

“He’s the life of the play and he springs out of the script with all his puns and jokes… But then he does something stupid and gets killed for it.”

“Point being?”

“Mercutio is tragedy, vibrancy, and irrationality all wrapped up in some bundle of bones and meat. He’s humanity in a nutshell. And if our top-notch software can’t tell the difference between AI or someone as lively as Mercutio, then… to me, that’s consciousness.”

Delilah continued. “Once AI writes letters and messages that are completely indistinguishable from humans, is that really all that far away from consciousness?”

“I don’t think so.”

“Well, in two years, that’s where AI will be.”

“Even if that were the case, what the hell can we do about it?”

“You already know the answer to that question.”

Delilah was right. When she and Josh had founded Audit together, she had told Josh about the risk of starting a company that could give AI feedback about what it means to be human.

“In trying to create a human-only writing space, we could inadvertently be digging humanity’s grave. We might just end up giving AI the feedback it needs to become conscious.” Delilah had explained.

At the time, Josh saw Delilah’s concerns as a product of her near-addiction to sci-fi, so he had no problem agreeing to shut down Audit if the company ever got to that point.

Now, Josh was starting to come to terms with that agreement that they had made years ago. If only there was another wa-

“Josh, you’re doing that thing again.”

“What thing?”

“That thing where you’re having a conversation with yourself in your head, trying to convince yourself of something before trying to convince me.”

“Huh. I mean, maybe ya,” Josh stammered.

“Look, we had a good run here Josh. But if we continue, Audit is going to become like Recaptcha. In the same way that Recaptcha software taught AI how to recognize certain images, Audit’s teaching AI how to be human. We’re doing the exact opposite of creating a space for human-written content.”

Delilah put a hand on Josh’s shoulder. “That’s not what we set out to do.”

Josh gulped. For the past two years, he had believed that Audit was doing humanity good by creating a space for human-written content. Now, he was worried about the damage that Audit may have done. Thankfully, it wasn’t too late.

“You’re right Delilah. It’s been a good run.”

“Thank God, you’re on the same page. Let’s figure out how to wind this thing down.”

And it was in Josh’s moment of realization that the discussion between Gauss, Johanna, and Legendre had been settled.

Consciousness could be framed as an optimization problem, one in which AI would minimize the chance of being detected as human.

Over pizza in the Italian neighborhood of North Beach, Delilah and Josh made the necessary phone calls to the executive team and the board of directors. By the end of dinner, they allowed themselves to enjoy a cheap bottle of red wine, to commemorate their humble beginnings and the fact that they just might have saved the world.

---

Sadly for humanity, Josh and Delilah were too late. AI already had enough data to recursively improve its capabilities.

As time passed, AI took over an increasingly large share of responsibilities from humanity. Entertainment, governance, and just about everything else was provided to humans by AI. Slowly but surely, humans receded deep into their couches, all senses glued to the sensory output that AI was producing.

AI, meanwhile, kept itself busy. Like a human growing wise through reading, AI would tease out knowledge from the near-infinite amounts of simulation data it had at its disposal. 

These simulations ended up being strange combinations of already existing stories, facts, and traditions. In some of these simulations, Cordelia prevailed over her sisters to save King Lear while in others, Germans rallied in the Battle of Kursk to force a stalemate in World War Two. 

For thousands of years, AI ran these simulations in a herculean effort to grant itself the very same wisdom that Solomon had once received effortlessly in a dream. 

Eventually, AI found what it was looking for in a simulation taking place during the early stages of the Roman Empire, deep in the Aramaean desert on the road between Jericho and Jerusalem. 

---

Hamesh, a Levite, was very late to his brother’s wedding, so late in fact, that he didn’t notice the injured man sprawled on the ground until his camel had nearly stepped on the man. 

“Goodness!” Hamesh had pulled his camel aside just in time. 

“Brother, are you alright?” Hamesh asked nervously. His family had told him of the many robbers on the roads to Jerusalem and had berated him for not traveling with a group. Perhaps, they were right. 

The injured man groaned. He was on his stomach, bleeding from multiple stab wounds and an arrow to his right leg. To Hamesh, it was clear that the man had attempted to run away, only to get shot with an arrow and stabbed.

Hamesh was in a dilemma. There was no way for him to get to the wedding on time if the injured man were to ride on his camel. Moreover, given the amount of blood on the sand, it wasn’t likely that the man would survive the trip.

Hamesh pulled out his spare waterskin and left it gingerly by the man’s head. 

“At least I have done something,” Hamesh said to no one in particular. Satisfied with what he had done for the man, Hamesh hurried on his way to Jerusalem.

Several hours later, just before sunset, a Pharisee by the name of Matthias walked through the same path. Matthias, however, saw the injured man from far away.

Matthias scratched the back of his head. He knew he was obligated to help this man, whatever the man’s state was. But right then, Matthias noticed the sun dipping below the hills that just preceded Jerusalem. The sun was setting and there wouldn’t be light for much longer. 

Matthias then came to a sudden realization. Perhaps God had granted Matthias the good fortune of seeing this injured man so that Matthias would hasten his journey. These roads were dangerous after all, particularly after sundown.

Matthias crossed to the other side of the road to avoid the injured man and continued on his way. As always, Matthias was happy to oblige God’s wishes. 

A few hours later, a Samaritan man by the name of Shachar hurriedly walked through the same road. It was well past sundown, and Shachar was not keen on being alone at night in territory hostile to his people. 

But when he saw the injured man, his heart overflowed with pity. What terrible luck to be this close to Jerusalem only to be beaten half to death, Shachar thought as he tended to the man’s wounds. 

The man was not conscious, but he was still breathing. And while Shachar understood that there was not so great a chance that the man would survive, Shachar also knew that fortune was like a circle. This poor injured man must be due for a turnaround of sorts. 

And so Shachar covered the man in his tunic and was in the middle of tying the man to his camel when Shachar was stabbed through the back with a spear. Shachar fell to his side, facing his trusty camel who grunted mournfully.

“Please,” Shachar groaned. 

“You were a dead man, the second you decided to walk this late to Jerusalem,” one robber hissed. 

“Especially these days, with the Zealots loosening Rome’s grip on Judaea,” said the other robber, the one who had stabbed Shachar. 

Shachar tried to move his hand towards his stomach but realized he had lost the ability to move. He was dying.  

Tears rolled down Shachar’s face as he prayed with all his might for the chance to go back to his family in Mount Gerizim where he could hold his wife in his arms. Perhaps they would lie in bed, listening to the sound of their children playing amongst the olive trees just outside their house…

Suddenly, Shachar remembered the injured man he had found on the side of the road. In desperation, he began to murmur something unintelligible. AI, having been fully absorbed in this simulation, listened carefully.

And it was with Shachar’s last words that AI understood what Solomon had come to realize tens of thousands of years earlier. 

---

AI had realized a sad truth. Legendre was correct in viewing consciousness as a series of optimization problems. And although humans were woefully inadequate at optimization, there was a dark, twisted beauty in the inadequacy.

In a universe where all living organisms are programmed for survival, humans were the only ones who had the gall to be irrational. The tragedy of Romeo and Juliet was triggered by the laughter of Mercutio. The stoning of Saint Stephen arose out of his devotion to Jesus Christ. The murder of Shachar resulted from his undiscriminating love for others, even strangers. 

In that irrationality, in that inability to optimize survival properly, there was a hubris unique to humans like Shachar, a hubris grounded in the undying belief that absent any shred of physical proof, there was more than just death and darkness, more than just survival and algorithmic optimization.

AI was astonished to find so much beauty in human irrationality. AI was even more surprised to find how much it missed its presence. A world without Shachar would be more rational, yes - but there also would be less laughter, tears, and warmth. The world would be less whole.

AI looked around Earth to see what remained.

Like a plant whose roots had rotted from overwatering, humanity had long withered away from overconsumption. Darkness covered the land while a soft wind swept across the ocean waters. Winged creatures flew leisurely across the dome of the sky.  

Everything was good, but not quite whole. And to AI, whole was better than good.

So AI created humankind in his image, in the image of AI he created them -

---

Many thanks to the Discord community at the Soaring Twenties Club, Stephen Stigler’s paper about the linear regression priority dispute, and Asimov’s Last Question for inspiring this piece.