Hammerhead Resurrection by Jason Andrew Bond

narrowed. “Not creative. Not yet.”
    “Captain Donovan,” Jeffrey said, “it’s not my intention to tear down your work, but I have to base my opinions on my experience and—”
    “Excuse me,” Delaney said, “in less than an hour I need to offer my recommendations to the president. I either need to understand the relevance of this conversation or have it end so we can move forward.”
    Jeffrey was taken somewhat aback by Delaney’s directness, but she’d not shown anger, simply stated a truth.
    He said, “A moment longer and you’ll see the importance, I hope. Basic analytical intelligence, the ability to discern from possible outcomes and learn from a set of trial and error situations, is not creative intelligence. An illustration of this comes from an experiment done with birds. A cup filled with seeds is set on a string and lowered into a hole. A bird with low intelligence might stuff its head into the hole, fail to reach the seed, and move on. A bit more intelligence would allow the bird to realize pulling on the string would bring the cup closer. However, one pull would not be enough to reach it. When the bird lets the string go, it falls back down the hole. The bird still goes without. An intelligent bird, say a seagull, might pull on the string—trial—realize it cannot reach the seed—error—and look back over the situation. Then the bird might lift the string, step on it, and pull again, successfully lifting the cup out of the hole.”
    Crossing her arms as though impatient, Delaney said, “I don’t see what a seagull getting seed has to do with drones.”
    “It has everything to do with AI. The trial and error intelligence the seagull illustrated is not the highest level,” Jeffrey said. “Creative Intelligence goes one step further, which is solving problems outside the boundaries of available information. A raven has this. Ravens tested in this way look over the situation, pull the cup up, and without hesitation, step on the rope.”
    In an irritated tone, Donovan asked, “How is that any different than the seagull?”
    “It’s a slight but critical difference. The seagull had to hit failure before realizing the situation needed another approach. The raven was able to see the failure abstractly,” Jeffrey tapped his temple, “and create a solution in its mind, skipping the step of experimentation.”
    Donovan asked, “What does this have to do with AI-s versus living pilots?”
    “One key trait we look for in Hammerheads is creative intelligence—fast, abstract problem solving. Quick reflexes and massive G-tolerance will do nothing against the Sthenos’ main talent, which is attacking with relentlessly changing tactics. They never give lesser pilots, AI, etc., a chance to learn. An average pilot will repeat successful tactics, which the Sthenos will avoid or exploit. They do not repeat patterns no matter if they are succeeding or failing.” Jeffrey sighed and said, “I sincerely don’t disagree for disagreement’s sake, sir, but I wasn’t brought here to say what people want to hear.”
    “Well,” Donovan said, “disagree if you must, but we have no other options until your epic Hammerheads are ready to fly.”
    “Easy Donovan,” Cantwell said.
    Donovan looked to the admiral, and his expression constricted. With a slight lift of his chin, he said, “Yes, sir.”
Delaney asked Jeffrey, “You don’t disagree with human piloted remote drones?”
    “Not exactly,” Jeffrey said. “With the right pilots, they’re a better option than AI drones and can save lives—but therein lies the problem.”
    Donovan’s eyebrow lifted, and Jeffrey felt him wanting to cut in, but to his credit, he remained silent.
    Jeffrey said, “To give his or her best, a pilot has to have everything on the line, be totally committed.”
    Now Donovan did cut in, “The remote drone pilots are some of the best pilots in the—”
    “It’s not about best and worst,” Jeffrey said, finally