The ELI Event B007R5LTNS

its acceptance at the Center, and had now managed the project for five years. E-L-I was an acronym for its full project name, Electroneural Learning Interface I. The name immediately became “Eli” to everyone except Wheeler, who insisted on calling it “E-L-One.” It was, after all, his project.
    Eli listened to Wheeler humming down the hall, the man he considered his creator, the man who had donated the living neural tissue that made him unique among artificially intelligent machines. Wheeler called him the first in a new generation of “test tube computers,” and Eli understood the reference.
    This computer knew it was unique in many ways, the most unusual of which was the use of living neurons in part of its circuitry. A huge machine, it was reminiscent of the room-sized computers of the 1950s. Its control panel, in fact, took up fully half of one wall of the underground lab which housed it. Much of the room was occupied by multiple banks of servers—originally eight banks of eight, now grown to sixteen banks of sixteen. At one end of the rectangular room was the computer’s bioneural module, a sealed aquarium-sized affair. It contained what was essentially an oversized Petri dish, a glass plate holding a set of circuits unlike any other on earth. In those circuits were hundreds of thousands of human neurons—Stephen Wheeler’s neurons—bonded together in a nourishing, self-replenishing organic soup and connected via a series of electrodes, radiating from the dish in all directions, to the rest of the machine’s internal components.
    Eli was the first of a truly new breed of self-designing, self-modifying computer, the next generation of biological neural networks. The more ordinary parts—the main memory, arithmetic logic, and central processing units—were assembled first, and the computing power of the original machine was then used to help determine how human neurons could be fused together as a working extension of the basic unit. Wheeler, as the project’s supervisor and primary designer, insisted on donating the nerve cells himself. When he was comfortable with the plan, he had checked himself into USC University Hospital and let them drill into his skull. The resulting tissue sample containing thousands of his neurons was then installed in and interfaced to the machine. As the displaced neurons began to interact with each other and their newly connected components, the network literally became alive.
    A brilliant computer scientist with more degrees than a thermometer, Dr. Stephen A. Wheeler was a relaxed thirty-two, brown over brown, and was considered something of a maverick, more comfortable in jeans and t-shirts than neckties and lab coats. He wore his hair a bit too long, and shaved only irregularly. In some ways, he resembled a young version of the stereotypical absent-minded professor, a resemblance with which he was entirely comfortable—he felt his mind was far too filled with things that really mattered to clutter it with trivial minutia. Hence the pocket recorder. He was never without it, and used it to keep track of the mundane, easily forgotten necessities of daily life, a category into which he lumped almost everything except E-L-One.
    The project was everything to Wheeler. Never married, not even close, he had devoted all his time and energy, his entire career, his every waking moment—and many of his sleeping ones—to E-L-One. He sometimes said he couldn’t remember ever wanting to do anything else, and meant it. His childhood, his education, his whole life up until the time he joined the Center were to Wheeler an extraneous and irrelevant blur, little more than a path he had to traverse, a journey he had to take just to get where he wanted to be. And where he wanted to be was right here, living, working, eating, sleeping, breathing nothing but E-L-One. It was his whole life, and he was completely okay with that.
    The soft click of Wheeler’s pocket recorder often reached Eli’s aural sensors. He wondered why a brilliant man like Wheeler needed such a primitive device. Then, in a small flash of insight, he realized that Wheeler must use it as offline data storage so that he could devote his primary memory resources to the most urgent tasks. He was satisfied with the deduction and stored the successful inference process for later heuristic analysis.
    Eli was pleased with many of his recent accomplishments. His neural network was of course bidirectional: it could not