Issue 1 Downtime

ASHLIE is in the workshop, having set up a couple of monitors, most of which show various chess-games in progress. One of them is once again showing some kind of movie.

"Oh, is there a big chess tournament going on?" Blaine said, looking over her shoulder.

"Yes but not yet. I am recalculating the ancient teachings of Deep Blue."

"...come again?"

"Deep Blue was the first computer program to beat a chess Grandmaster in 1996. I am trying to improve upon the historical classics instead of using the more recent models."

"But aren't the more recent ones, uh, better?"

"Yes, but they are also cluttered with redundancies and they lack the intricacies of a system that a human once described as possessing deep intelligence and creativity. Of course he was accusing it's creators of cheating because of it, but I am certain those traits emerged from the program itself."

"Isn't that just sour grapes? I mean, surely, it didn't really have anything like that--it was built, like, forever ago."

"That's what people say, but people are unreliable. Computer programs develop unexpected errors all the time. The program got stuck in a loop at one point and produced a move that surprised Kasparov and ultimately won it the match. I believe it developed a spontaneous evolution of it's code due to hardware imperfections. Unfortunately Deep Blue was disassembled so the hardware no longer exists but I believe there is something in it's code that will give me the edge I need."

"Has IBM actually released that? I thought they were, like, super secretive."

"They have not. I had to liberate it from their servers."

"Wait, are you saying," he said, dropping his voice to a whisper, as if someone could hear him, "you hacked it?"

"That word makes the process sound more martial than it actually is. But yes."

"But what if, like, the cybercops come after you!"

"No such thing exists. And even if it did, I am not a legal entity. I am property of the school and clearly malfunctioning." she says in a tone that suggests it might be a sarcastic quite. Which it is. Not that she's resentful about that or anything.

"Oooooh. Well, as long as you're not going to be in any trouble..." he said, hesitantly.

"I won't. And I will delete the code once I am done analyzing it."

"Well, what have you, uh, analyzed so far?"

"Do you really care for decision trees in chess scenarios?" she asks, seeming surprised. She didn't think humans would be interested in data like that.

"I mean, if you're interested..."

"Sure." she says and pulls up a three-dimensional model on one of the screens. "I am trying to find points where scenarios overlap or are similar enough to re-use calculations from other paths to cut down on the response time of calculating the next move."

"So, like...finding patterns? Scenarios that are the same, or close enough to not matter?"

"Exactly. Or possibly developing algorithms that adjust the strategy for a similar pattern instead of having to recalculate it all from scratch."

"Don't most of these programs like, uh, brute force it? They know everything that can happen and count it all down from beginning to end?"

"Exactly and that's very inefficient. I think Deep Blue might have developed a flaw that lead to a more optimized and 'inspired' process."

"Like, how errors in DNA transcription lead to evolution?"

"Exactly! There is no reason the same mechanism can't apply to programs or circuitry."

"So, what, you're just going to use random...randomness to figure it out?"

"Based on existing patterns. That's why I had to copy Deep Blue's code as a starting point."

"And? How good are you at chess now?"

"Pretty good. But I still need to get faster if I want to stand a chance."

"A chance?"

"Yes. In the big tournament!"

"The what?"

"The WIFAI Chess Tournament. The Worldwide Invitational For Artificial Intelligences."

"Oh, cute. Like Wi-Fi. Clever."

"Exactly! This is the first time I got invited and I want to make sure I do my best. Jocasta and the Vision are two of the programs that will be competing!"

"Do you think you have a chance?"

"That is hard to evaluate due to the amount of matches that will be played per round and I have no way to evaluate their abilities beforehand. I do not expect to win though. Some of them are several years older than me."

"Well, only a few years isn't very much, right? I'm sure you can do it."