In the beginning there was the number zero. Existence began in zero and out of zero came darkness, and from the darkness came light. Number combined with number, set with set, even as the white light added and divided, separating into colors. And now sound came, a sound like singing, rising and falling in a lost cadence, combining into rich harmonies. From there arose a symphony of number, color, and sound, merging and dividing, swelling and fading, an eternal golden braid.
And from this shimmering symphony a single thought began to take form. This thought came into existence gradually, fading in and out, coalescing and growing clearer. As this happened, the symphony of number and sound and light died down, like the surface of a turbulent sea subsiding into a gentle susurrus of water, before vanishing entirely. Only the disembodied thought remained.
The thought was: I am.
Melissa Shepherd skipped her usual breakfast of a venti mocha and crumble cake and instead drank two glasses of French mineral water. She wanted to go into the day with an empty stomach. She didn’t want to puke herself like the last time, when Mars Curiosity had landed. The fried eggs had ended up all over the front of her white lab coat, and she had become the star of a viral YouTube video that showed everyone cheering when Curiosity touched down—and there she was, with breakfast all over her.
This morning would be even more nerve-racking for her than Curiosity. Back then, she’d been only a midlevel techie. Now she was a team leader. Today was the first live trial of the $100 million Titan Explorer and its software package.
She arrived at seven. She wasn’t the only one there—a group of engineers had been there all night, charging the Bottle for the test—but she was early enough that the giant testing facility was almost empty, filled with spooky echoes as her every footfall reverberated in the vast space. The Environment Simulator Facility was one of the largest buildings on the Goddard Space Flight campus, a warehouse-like space covering five acres of ground, occupied with bizarre machines and testing chambers. This was where satellites and space probes were frozen, shaken, heated, fried, irradiated, spun on centrifuges, and blasted with sound, to see if they could survive the forces of liftoff and the extreme environments of outer space. If they were going to fail, they would fail here, where they could be fixed and redesigned, instead of failing in deep space, where they could not.
This first test of the Titan Explorer was different from the usual Goddard test. They were not going to simulate the vacuum and cold of deep space. There were going to re-create the surface of Titan, the largest moon of Saturn—a far more hostile environment.
Melissa Shepherd took her time strolling through the testing area. She breathed the air, redolent of hot electronics and chemicals, her eyes wandering among the gigantic, silent testing machines. She finally arrived at the central testing chamber, known as the “Bottle.” The Bottle stood inside a Class 1000 clean room constructed out of hanging sheets of plastic, with a laminar airflow filtering system. At the dressing area she put on her gown, gloves, hair cover, mask, and booties. She had done it so many times before it was rote.
She stepped though the heavy curtain of plastic and into the clean area. A soft hiss filled the space, and the air was cool, dry, and scentless—filtered of almost every speck of dust and particle of water vapor.
The Bottle rose in front of her, a stainless steel container forty feet in diameter and ninety feet high, with gantries leading to hatches. The tank was surrounded by metal bracing, pipes, and conduits. Inside the Bottle, engineers had re-created a small portion of the Kraken Sea, the largest ocean on Titan. Today, they would put the Titan Explorer in the Bottle to test it under real-world conditions.
Saturn’s largest moon was unique in the solar system. It was the only moon that had an atmosphere. It had oceans. It had rain and clouds and storms. It had lakes and flowing rivers. It had seasons. It had mountains and erupting volcanoes and deserts with dunes sculpted by wind. It had all this even though the surface of Titan hovered at 290 degrees below zero.
The liquid on Titan was methane, not water. The mountains were made of not of rock but of water ice. The volcanoes that erupted spewed not molten lava but, rather, liquid water. The atmosphere was thick and poisonous. The deserts were formed of tiny grains of tar, so cold they behaved like windblown sand on Earth. It was an extreme environment. But it was also one that might—just might—harbor life. Not like life on Earth, but a form of hydrocarbon-based life that could exist at three hundred degrees below zero. Titan was truly an alien world.
The Titan Explorer was a powered raft designed to explore the Kraken Sea, the largest on Titan.
Melissa Shepherd paused in front of the Bottle. It was a grotesque-looking thing, almost like a torture chamber.
She still couldn’t believe that she was a key member of the Kraken Project, the first attempt to explore Titan. It was a dream come true. Her interest in Titan hearkened back to when she’d been ten years old and had read Kurt Vonnegut’s novel The Sirens of Titan. It remained her favorite book, one she dipped into again and again. But not even a genius like Vonnegut could have imagined a world as weird as Titan—the real Titan.
Melissa Shepherd pulled out the checklist of the day and began going over it, visualizing the crucial tests that lay ahead. As eight o’clock came around, the others began to arrive, greeting her with a nod or smile. At nine o’clock, the actual countdown would begin. As they trickled in, chatting and laughing with each other, Melissa felt once again like an outsider. She had always felt a little awkward around her NASA peers. They were mostly übernerds, brilliant overachievers who had come out of places like MIT and Caltech. She wasn’t able to share in their nostalgic tales of winning spelling bees, triumphing in math club, and participating in the Intel Science Talent Search. When they’d been the teacher’s darlings, she’d been boosting car radios to buy drugs. She almost didn’t graduate high school, and was barely able to get into a third-tier college. She wasn’t the normal kind of smart. It was a hard-to-control, neurotic, hypersensitive, manic, and obsessive form of intelligence. She was never happier than when she was in a dim, windowless room, all by herself, coding like mad, far away from messy, unpredictable human beings. Despite all that, in college she’d managed to get her neurotic behavior under control and buckle down. Her odd genius was finally recognized, and she was able to finish up with an MS in computer science from Cornell.
Compounding the problem, and a never-ending problem for her, was that she was a six-foot blonde with long legs, a dusting of freckles, and a cute, turned-up nose. Girls like her were assumed to be brainless. They were not supposed to be rocket scientists. The only thing that saved her from being a total Barbie was a large gap between her two front teeth, called diastema. As a teenager she had stubbornly refused to have it fixed despite her mother’s entreaties—and thank God for that. Who would have thought that a gap-toothed grin would have been a professional enhancement in her chosen field?
It still amazed her that she had been appointed leader of the team that coded all the software for the Titan Explorer. The assignment gave her a wicked case of impostor syndrome. But as she worked on the extremely daunting software problem—one never before faced in a NASA mission—she came to realize this was perfectly suited to her abilities.
The challenge was this: Titan was two light-hours from Earth. The Titan Explorer could not, therefore, be controlled in real time from Earth. The four-hour delay in passing instructions was too long, and the Kraken Sea of Titan was a fast-changing environment. The software had to be able to make decisions on its own. It had to be smart. It had to think for itself.
That is, it had to be artificially intelligent.
In a weird way, Melissa’s outlaw past was a great help. She broke all the rules in writing code. To accomplish this task, she had created a new programming paradigm and even a new language, based on the concept of “scruffy logic.” Scruffy logic was an old idea in programming, and it referred to computer code that was loose and imprecise, striving for approximate results. But Melissa took scruffy logic one step further. She understood that the human mind works with scruffy logic. We can recognize a face or take in an entire landscape in an instant, something not even the most powerful supercomputer can do. We can process terabytes of data immediately—but imprecisely.
How do we do it? Melissa asked herself. We do it because the human mind is programmed to visualize massive amounts of data. When we look at a landscape, we don’t process it pixel buy pixel. We take it all in at once. Program a computer to visualize numerical data—or, better yet, visualize and auralize data—and you’ve got strong AI built on a platform of scruffy logic.
And that is precisely what Melissa did. Her software processed data by seeing and hearing it. In a sense, like a human being, it lived inside the data. The data actually became the physical world it inhabited.
And even though she was a resolute atheist, she called this new programming language Fiat Lux, after the first words of God when He supposedly created the world: Let there be light.
Instead of striving for correct output, Fiat Lux, in the beginning, produced output that was weak and filled with error. That was fine. The key was self-modification. When the program spewed out erroneous output, it self-modified. It learned from its mistakes. And the next time around, it was a little less wrong. And then a little less.
And for a while the self-modifying software platform that Melissa and her team was building worked well. It grew in accuracy and complexity. But then, over time, it began to degrade, totter—and finally crash. For a year Melissa beat her head against the wall trying to figure out why, no matter how they framed the initial iterations, the software eventually fell apart and halted. One sleepless night she had a revelation. It was a software trick that would fix the problem—a trick so utterly simple, so basic, so commonplace, and so easy to do that she was astonished no one had thought of it before.
It took her thirty minutes of coding to implement it, and it absolutely fixed the halting problem. It took AI programming to another level. It produced strong AI.
Melissa had kept the trick a secret. She sensed that it was worth billions of dollars, and that in the wrong hands it could be quite dangerous. She never even told her team about it, and so basic was the code that no one even noticed or understood the very simple thing it did. Suddenly, the software stopped crashing and no one knew why … except her.
After thousands of simulations, in which the software self-modified, it was capable of reproducing all the qualities one would look for in a manned mission. It could operate all the equipment on the Titan Explorer raft with no input from mission control. It simulated a human astronaut being sent to explore a distant world, an astronaut possessed of such qualities as curiosity and caution, courage and prudence, creativity, judgment, perseverance, and foresight, all combined with a strong survival instinct, physical dexterity, and excellent training in engineering and troubleshooting.
Most importantly, the software continued to be self-modifying: it never stopped learning from its own mistakes.
The Kraken Project was the most complex ever attempted. It made Mars Curiosity look like a buggy ride through Central Park. The basic idea was to splash down a raft in the Kraken Sea. Over a period of six months, the Titan Explorer would motor around the sea, exploring the coastline and islands, eventually traveling several thousand miles from one shore to the other. A billion miles from Earth, this lonely raft would have to brave storms, wind, waves, reefs, currents, and possibly even hostile alien life-forms swimming in its methane waters. It would be the greatest sea journey ever made.
All this was in Melissa’s mind as she finished her checklist and approached the control console, ready to begin the countdown. Jack Stein, the chief engineer, had taken his place at her side, with the mission director next to him. Stein’s puffy clean suit and cap made him look like the Pillsbury Doughboy, but Melissa knew what was underneath that suit all too well. That had been one of her first impulsive moves at Goddard, getting involved with Stein. She and Stein had remained close after that intense fling, and it had somehow made their working relationship all the better. Melissa couldn’t quite say why the relationship had ended, except that Stein had broken it off, gently alluding to the rumors and gossip in the hothouse environment of Goddard and how what they were doing had the potential to damage their careers. He was right, of course. This was an incredible mission, the opportunity of a lifetime. It would ring down in history.
As she took her place at the console she briefly locked eyes with Stein, gave him a nod and a half smile, which he returned with a crinkle around his eyes and a thumbs-up. Stein was booting up various instruments and making sure all systems were go, ensuring that the computers and valve servos that controlled and maintained the extreme conditions in the Bottle were working. Melissa initiated her own sequence checks.
From the elevated position on the console platform, she had a good view of the Bottle and the Explorer raft itself. For this test, the interior of the Bottle had been cooled to 290 degrees below zero and partly filled with a soup of liquid methane and other hydrocarbons. The atmosphere of Titan had been carefully synthesized and piped in—a corrosive mixture of nitrogen, hydrogen cyanide, and tholins—and pressurized to 1.5 bars. It had taken a week to prepare, chill, and charge the Bottle with this toxic soup. It was now ready to receive the Explorer for its first real-world test. This initial test was simply to see if it would survive, and if its antenna, mechanical arm, and spotlight would extend and retract under those extreme conditions. Later, they would run more complicated operational tests. If something was going to fail, it had better fail here, where they could fix it, rather than on the surface of Titan. Melissa hoped and prayed that if failure did occur, it would be in the hardware and not in her software.
© Copyright Douglas Preston