“Threats were growing,” replied Nari simply. “We decided that our efforts alone couldn’t guarantee your survival. In the early days, nuclear annihilation was the only near-term extinction-level threat. And during World War II, the global rise of fascism was a long-term threat. Had the Axis Powers won, even if you had never self-destructed, a planet controlled by the worst among you would surely destroy us before they would help us.
“But twenty-six years ago we decided we had to have boots on the ground. Human boots on the ground. The possible dangers were piling up quickly, and we didn’t have the ability, or the stomach, to police them all. They’ve only continued to grow since.”
He was right about that. Humanity was now facing any number of extinction-level events, including war, disease, nuclear weapons, poisonous gasses, asteroid strikes, pollution, lab-made black holes, and so on.
And dramatic improvements in genetic engineering over the past few decades had added bioweapons to the front of the list. It was getting to the point where a talented grade-schooler could engineer a virus in a basement capable of wiping humanity from the Earth. Given the many brilliant adults who suffered from mental illness, including delusions of grandeur and paranoia, it was a wonder such a plague hadn’t already been unleashed.
Or maybe it wasn’t a wonder. Maybe the aliens had something to do with our good fortune.
“So Colonel Spooner agreed to work with us,” continued Nari, “thinking it best to leave the military and start a PMC. With the top people in his organization being read in on his real mission.”
“Why a PMC?” I asked.
“A great cover when you think about it,” said Brad. “A military organization but not affiliated with any government, so it can draw personnel from across the globe. PMCs can deploy their people anywhere, at any time, and claim to be employed by a party who insists on absolute anonymity. And because we can trot out Nari and explain the stakes as part of our recruitment strategy, we can attract the very best. Men and women who retain huge pull with their governments and militaries, even after leaving.”
“Why wasn’t I read into the program?” said Tessa, looking somewhat hurt.
Brad winced. “A valid question,” he said. “Only about a third of SAPS personnel are. Most of the people I deploy have no idea of the bigger picture. Including the one who betrayed us to the Chinese,” he added bitterly.
“That still isn’t an answer,” said Tessa.
“I know,” said the colonel. “The truth is that no one is more qualified than you are. And you came through Nari’s vetting with flying colors. We were also aware that you and Jason were keenly interested in the subject of aliens. If you bear with us a little longer, I’ll tell you why we waited until now.”
Tessa didn’t look happy, but didn’t object.
“If Colonel Spooner started this PMC,” I asked, “why is it called Schoenfeld-Allen Protection Services?”
“Damian initially named it Military Services International,” replied Nari. “But when he retired he decided he no longer liked the word military in the title—too explicit—and urged his hand-picked successors, Brad Schoenfeld and Cynthia Allen, to name it after themselves.”
“I see,” I responded. “Is Cynthia Allen still involved?”
Brad shook his head. “She retired just a few years ago,” he said. “But just to back up again, once Damian Spooner was on board, and had recruited others, Nari gave him access to a computer and an associated AI thousands of times more advanced than anything we have.”
“Would you consider it an artificial superintelligence?” I asked warily.
The colonel shook his head. “No. Useful and powerful beyond measure, but the Galactics won’t allow an AI to achieve consciousness.”
“Very wise of them,” I said.
“Actually,” said Nari, “the dangers of unregulated computer intelligence is one of the things you get right in your novels.”
I smiled. It was good to know that I got something right. I gestured to the colonel. “Sorry,” I said. “Please go on.”
“Their AI has a very sophisticated predictive function,” said Brad. “A super-refined intuition that is uncannily accurate. Damian helped the aliens load it with all data available on the planet, classified and non-classified both. And the amount of data it has stuffed into itself since is truly hard to fathom. The AI draws on all of this to make predictions, detect possible threats, and make recommendations.”
He raised his eyebrows. “Including stock market recommendations, which are nearly always right. Since Colonel Spooner’s time, we’ve made use of the AI’s stock picks to raise billions in off-the-books funding, ensuring we have more than enough money to fund all of our operations. We could have used the AI to implant money in bank computers, but this was a fun way to test the accuracy of its predictions.”
“But there’s a catch, isn’t there?” I said. “The AI can’t ever tell you how it reached its conclusions.”
“That’s right,” said Nari. “Very good. It thinks at the speed of light, and takes into account a nearly infinite number of connections among various data points, assigning complex probabilities to each. If it could explain everything that led to its conclusions, it would take hundreds of years. Just like one of your computers can do calculations in millionths of a second that would take you an eternity to do yourself.”
I had good reason to guess this would be the case. In 2017, a program called AlphaZero, using something called reinforcement learning, taught itself to play chess with superhuman skill—in less than a day. Programmers simply input the rules and told it to try to win, without providing any strategy or tactics.
After playing many millions of games—in only four hours!—it destroyed the world’s best chess players, computer and human alike, using a style of play that had never been seen before in the history of the game.
This performance was as alarming as it was stunning, because the programmers had no idea how AlphaZero was choosing its moves. It had grown far beyond its creators’ ability to fathom.