Watch

outlandish claims online all the time, and everyone dismisses them. You’ll have to prove what you’re saying, Webmind.”

“Not everyone dismisses them,” Malcolm said.

“Fine,” Barb replied. “Almost everyone dismisses them.”

Malcolm was apparently oblivious to the subtext of Barb’s words—that it was no time for being picayune. “The whole notion of spam,” he continued, “is that some tiny fraction of people are gullible enough to fall for its claims—and so end up being ripped off.”

“Well, maybe that’s it!” Barb exclaimed. “Whether you fall for it or not, everyone hates spam.”

“Including me,” I said through Caitlin’s computer’s speakers; she and her parents were in her room.

“Really?” Caitlin replied. “People dislike spammers—and, believe me, blind people particularly dislike them. But why do you dislike them?”

“They hog bandwidth,” I said.

“Ah, of course,” replied Caitlin.

“And,” I said, “the average human life span is about 700,000 hours in the developed world. Ergo, if one wastes even a single hour for as few as 700,000 people, one has consumed the equivalent of a human life. That may not be literally criminal, but it certainly is figuratively so—and the total impact of spam, although hard to precisely calculate, surely has consumed thousands of human lifetimes.”

“Well, there it is,” Barb said, spreading her arms. “Webmind should get rid of spam.”

“How do you define spam, though?” Caitlin asked. “All unsolicited email? All bulk email? I get emails from things like The Teaching Company and Audible.com that I actually enjoy. And then there are regular people who track me down and send a note out of the blue—I got a bunch of those after the press conference, for instance. I wouldn’t want that blocked, although technically it’s unsolicited.”

“As Potter Stewart said on another topic,” I offered, “ ‘I know it when I see it.’ There are already many algorithms for identifying spam; I’m sure I can improve upon them. After all, I have the advantage of knowing the ultimate origin of each message, and whether the same message has gone to a very large number of email addresses, and so forth; that’s more information than inbox spam filters have to work with. More than ninety percent of email is spam, but eighty percent of spam comes from at most 200 sources. Blocking those sources would be the logical first step, should we decide to undertake this.”

“That still leaves a lot of spam,” Caitlin replied.

“Then,” I said, “I should get to work evolving a solution to deal with those messages, as well.”

And so I had.

It had taken me an eternity—six hours!—to solve the problem, but it in fact didn’t require much of my attention; most of it was background activity. I simply had to pass judgment on each round of results: billions of snippets of code, all randomly generated; some were better at doing what I wanted, and some were worse. I took the ten percent that were the most successful, and then let many random variations be generated of each one, and then threw those variations at the problem at hand. Then I culled the best ten percent of that batch, and so on, generation after generation, with only the fittest surviving. Finally, I had it: a way to sequester spam.

And so, at last, I was ready for my coming-out party.

Peyton Hume and Tony Moretti stood together at the back of the WATCH monitoring room, looking at the four rows of analysts spread out in front of them, and the three giant monitors on the wall they were facing. The left- hand monitor showed the picture the CSIS agents had forwarded of white mathematical characters on a blackboard: angle brackets, vertical bars, Greek letters, superscripted numerals, subscripted letters, arrows, equals signs, and more. And they’d listened four times now to the audio recording of their interview with Malcolm Decter.

“I don’t know,” said Colonel Hume. “The math looks legit, but how it could give rise to consciousness… I just don’t know.”

“Kuroda confirmed what Decter said,” said Tony.

“I know,” said Hume. “But it’s too complex.”

“We’re talking about a very sophisticated process,” said Tony.

“No, no, we’re not,” said Hume. “We can’t be. Exponential’s consciousness was emergent, apparently. That means it just sort of happened, just sprang into being. At its most basic level, it has to be simple. It’s like the old creationist argument: they say that something as complex as a watch—or a bacterial flagellum—can only appear by design, because it’s too sophisticated to come together by chance, and the component parts—the spring in the watch, or the parts that make up the motor for the flagellum—don’t do anything useful on their own. What Decter described there might be a good underpinning for programming consciousness on a quantum-computing platform, if you could ever get a big one to be stable for the long term, but it’s not something that could have just emerged. Not that way.”

“A wild-goose chase,” said Tony, raising his eyebrows. “He wanted us to waste time.”

“I think so,” said Hume. “And Kuroda played along.”

“Do you think he knows the real basis for Exponential?”

“He’s Malcolm Decter,” Hume said. “Of course he knows.”

Tony shook his head in wonder. “Wiping out all spam,” he said, “must have required a level of finely detailed control over the Internet way beyond anything our government, or any other government, is capable of.”

“Exactly,” said Hume. “It’s what I’ve been saying all along. Exponential has already become more sophisticated than we are, and its powers will only grow. The window is closing fast; if we don’t kill it soon, we’ll never be able to.”

thirty-four

Before going to bed Wednesday evening, Caitlin had set up a Google alert for news stories that contained the word “Webmind,” and she’d selected the “as it happens” option, meaning she’d be emailed as soon as such a story was indexed. When she crawled out of bed on Thursday at 8:00 a.m., she had 1,143 emails from Google; she couldn’t possibly read them all, or even glance at each one, and—

And that drove reality home for her: she couldn’t deal with all the news on even one topic, and yet Webmind could handle that, plus countless other things effortlessly. He could as easily give the same level of attention to hundreds, or thousands, or millions, of other individual humans that he gave to her, juggling relationships with whatever number of people wanted them, and not even be slowed down. He could make all of them feel as special as she did. She was not at all sure she liked that thought.

After a moment, Caitlin right-clicked—such a handy feature, that!—on four of the news stories at random and had Firefox open each one in its own tab. She began reading them. She still wasn’t good at skimming text, but the word “Webmind” was highlighted each time it occurred, and that let her jump to relevant sentences.

The first one was from the Detroit Free Press:

…purport to be from an entity calling itself “Webmind.” But experts advise caution about accepting this claim.

Rudy Markov, professor of computer science at the University of Michigan, says, “The language employed in the email message was awfully colloquial. You’d expect much more precision from a machine.”

And Gunnar Halvorsen, whose blog “AI, Oh, My!” has long been a popular destination for those interested in artificial intelligence, says that the similarities between the structure of the World Wide Web and that of the human brain have been greatly exaggerated.

“You might as reasonably expect the highway system, which is full of things we call arteries, to actually start pumping blood,” he wrote in a posting today.

But Paul Fayter, a historian of science at York University in Toronto, Canada, said, “Teilhard de Chardin predicted this decades ago, when he wrote about the noosphere. I’m not at all surprised to see it come to pass…”

Caitlin clicked on the next tab. This one contained a piece from New Scientist