Checkers grandmasters have long suspected that perfect play would result in a draw, but until now, there has been no definitive proof. The first checkers-playing computer was created in 1963 by the artificial intelligence pioneer Arthur Samuel; the computer managed to win a single game against a human.

In 1989, Jonathan Schaeffer, who now heads the computer science department at Alberta, created Chinook with the aim of marshaling parallel processing and lots of storage to take on the world’s best players. In 1990, Chinook became good enough to enter the checkers World Championships, and in 1992, it faced off against the world champion — and the best checkers player who ever lived — Marion Tinsley. Tinsley narrowly defeated Chinook. Then, in 1994, the pair had a rematch, but Tinsley took ill and withdrew in the middle of the game. He died of pancreatic cancer a short while later.

“The unfinished Tinsley match left the question unanswered as to who was the better player,” Schaeffer and his colleagues write in this week’s issue of the journal Science, where their paper is published. But now the answer is clear: “As great as Tinsley was, he occasionally made losing oversights — he was human after all,” they say. Chinook will not make mistakes, and thus becomes the greatest checkers player of all time.

The research makes checkers “the most challenging popular game to be solved to date, roughly one million times more complex that Connect Four,” which was solved in 1989 (if two players play Connect Four perfectly, the first player will always either win or draw).

Their work also highlights the utility of raw computing power in intense artificial intelligence applications. In the early days of A.I. research, Schaeffer and his colleagues note, scientists often tried to make computers mimic human thought. But this approach led to difficulties, and they found that “human-like strategies are not necessarily the best computational strategies.”

A better method for solving complex tasks like checkers, A.I. theorists discovered, was “brute force” — rather than trying to master human strategies, computers would rely on “limited knowledge” of the specifics of the game, and instead use superior processing power to search and analyze all possible moves. That’s the approach taken by chess-playing computers — such as IBM’s Deep Blue, which beat champion Gary Kasparov in 1997 — and it’s also what Chinook does with checkers.

That their machine managed to solve the game, the researchers say, “provides compelling evidence of the power of limited-knowledge approaches to artificial intelligence.” This method will become even more powerful as computers themselves get faster and cheaper, they note. But solving games as complex as chess is still far off.

“Checkers has roughly the square root of the number of positions in chess (somewhere in the 10⁴⁰-10⁵⁰ range),” they write. “Given the effort required to solve checkers, chess will remain unsolved for a long time, barring the invention of new technology.” But disk-flipping game othello — solving that is possible, they say. The effort “will require considerably more resources than were needed to solve checkers,” but soon we’ll have an unbeatable othello-playing machine, too.

You can play against Chinook here.

But I can laugh at all of this now because I am off the roster. I am no longer “out there,” as they say. And I didn’t have to lower my standards or search outside my geographic region either. What I had to do was stop searching and start designing. That’s right. I designed my boyfriend. I’m a busy woman. I don’t have time for the Toms, Dicks, and Harrys the world keeps throwing at me.

Enter AI4U, top-of-the-line virtual-companion designers. No, they’re not cheap, but get real, they’re custom-designing your boyfriend. If it’s cheap I’m not interested. Granted, he’s a Web-based AI, not a flesh-and-blood man. So what? This isn’t about sex and anyway, the physical part of a relationship always fades eventually.

The design process is easy. First step: Pick a physical template. A youth squandered on Monty Python reruns left me with a full-blown kink for English guys, so I chose a template called “Nigel” — think Michael Palin crossed with Laurence Olivier. Then, to assure he didn’t look overdesigned, I clicked the “random factor” option to introduce “lifelike imperfections.”

As for Mr. Dreamboat’s personality, I had two options: I could allow AI4U to mine my Web habits, construct a psychological profile, and design my boyfriend’s personality to match. Or I could tell them in one hundred words or less exactly what I wanted. I chose the latter. I’m no privacy freak, but I didn’t want someone spying on my subconscious. Plus, when it comes to men, I know what I want. I don’t need some faceless software shrink hypothesizing about it.

I began with a firm list of no-nos culled from the rogues’ gallery of losers I’d dated over the years. Anyone bossy, intolerant, macho, repetitive, nosy, bookish, vain, foppish, anal, whiny, bipolar, fickle, sexist, nihilistic, or judgmental need not apply.

But I didn’t want Mr. Dreamboat to be defined by negatives, so I dredged the muck of my romantic archives for desirable traits. They were scant. There was Peter’s reliability. He said eight-fifteen, he meant eight-fifteen. James, despite the love handles and a wife in Greenwich, had initiative up the wazoo. Then there was Billy Sebert, who made me a papier-mâché model of his heart in sixth grade. That was sweet.

So on the plus side I had reliable, initiative-taking, and good with papier- mâché. That felt slim, so I added quick-witted, fun-loving, and emotionally balanced. For good measure I threw in the ability to rhyme at will, a passion for Shakespeare and an inexplicable love of the color orange. Why not, right? When I hit Send, a pop-up told me I’d hear from Mr. Dreamboat in forty-eight hours.

Exactly forty-eight hours later, I got the following e-mail:

Dear Lucy:

I hope you are well. If you’re free Thursday night, I’d love to show you around my neighborhood. Just goggle in to the following link — Pritchard_Booker.ai4u.com.

Cheers,

Pritchard

Reliable: check.

So Thursday night rolled around. I donned my favorite Janny Renoir suit — off-white, three-piece, custom-tailored to fit me like a glove. I slid into my almost-but-not-quite-obsolete RingletGloves — Titanium three-knucklers with a faux copper finish. And finally, I popped on my flash new UltraReality Goggles. Sharper Image $3,000. Thank you very much. For an extra $2,000, I could have had the matching UltraSensory full body glove, but at that point in time, I was in no way ready for a pervsuit. The very idea of virtual sex gave me the willies.

My pothead friend, Marla, is always telling me what a waste of money all this swank geekgear is, that I should be saving up to go neural like her. She’s even working a third job to buy herself the operation. I’ve always said the girl would turn herself into ones and zeroes if she could.

Once I’m goggled and gloved, I get comfortable on my white leather sofa, gog into my homesite, which is an exact replica of my living room, and give my avatar a good once-over in the virtual mirror. I dress her in the same suit I’m wearing, then try out a couple of different hairstyles. Nothing too extreme. I want to appear stylish, not desperate. I go for the mod bob. Ash blonde. Other than that (plus a little “help” in the chest region) I don’t falsify my avatar. What’s the point? What do I get out of pretending I’m a supermodel? The only person I need to impress tonight is designed to love me exactly as I am.

A cluster of icons for my favorite sites hover like thought balloons above my virtual desk. Among them is Pritchard’s link. I aim my ringletted right pointer finger at it and wink. It lights up in neon blue for a second, then everything goes black.

A moment later, I’m sitting on a bench in some swanky urban neighborhood with people criss-crossing the sidewalk in front of me. Avatars or scenery, I can’t tell. The taxis are huge and black and the sky is that dark steel blue you don’t find in the U.S. unless it’s about to rain. On the corner to my left is the giveaway: a person-sized red phone booth. London.

Across the street I notice a guy on a bench sneaking looks at me over the top of a newspaper. Longish dark hair, faded jeans, a camel tweed suit jacket and a muted orange shirt. When he sees me looking at him, he puts down the newspaper and heads toward me. There’s a four-lane street between us, which gives me plenty of time to give him the old up and down. Tall: check. Slim: check. When he’s about two-thirds of the way, I make out the details of his face: high cheekbones, a large nose angled slightly to the left and a wide goofy smile. His teeth are perfect, except for a twisted left incisor.

When he makes it onto the sidewalk, he offers his hand, looks me right in the eye and says, “Lucy?”

The “u” sound is deliciously English.

Much as I’d like to report that I took his hand, smiled confidently and said, “Why Pritchard, what a pleasure to meet you,” I cannot. I’m stricken, mesmerized, hypnotized by his eyes. They’re a shade of green I’ve never seen before. Not quite emerald, but brighter than hazel, with tiny flecks of black in them. They’re the only otherworldly feature in his achingly human rendering. The man is stunning. Not in any predictable way. He’s no movie star. He’s the guy in your Renaissance Lit class. You know the one: sits alone in the back, shoots you tortured glances you can’t interpret.

Well done, AI4U.

“Let’s take a walk,” he says. Then he lays this nuclear smile on me. I swear he could melt glaciers with that thing. So I take his arm and now we’re walking through virtual London. For reasons I can’t yet grasp, I’m too nervous to look at Pritchard, so I take in the sights instead. It’s a topnotch ‘scape, the chewing-gum-free sidewalks and reverse-flowing traffic utterly convincing. Even the snippets of dialogue from passersby are authentically English.

“I chose London,” Pritchard says, “because I thought you’d like it.”

“Yeah,” I say, a fountain of eloquence. “I guess I am a bit of an anglophile.”

He smiles darkly and says, “So I presumed.”

That’s when it hits me. It hits me like a slap in the face. This is no date. Pritchard is not some helpful, accidentally gorgeous Brit playing impromptu tour guide to my lost American tourist. He’s a gigolo. He’s an artificial gigolo and I’m a sad, thirtysomething spinster.

Now before you rain down a storm of duhs on me, understand that this is my very first intimate encounter with a humanoid AI. Sure I’ve interfaced with animated bots and had conversations with soft agents with human voices before, but I’ve never, you know, dated one. I have no analog for this. Somehow in the rush to design the perfect man, I forgot to anticipate what the actual date would be like. Now that it’s happening in real time, it feels dirty. And I don’t mean dirty in a good way.

To avoid revealing this inconvenient rush of squeamishness, I decide to clam up and let Pritchard walk me around. He tells me some quasi-interesting factoids about London’s architecture, then takes me to a public square where a handful of college students are mugging their way through a scene from “A Midsummer Night’s Dream.”

I manage a full sentence: “Do you like Shakespeare?”

Of course he likes Shakespeare. I designed him to like Shakespeare.

Pritchard laughs and says, “Is that supposed to be Shakespeare?”

I try to follow up with a joke. Something about Shakespeare turning over in his grave, but it’s pitiful. I’m in terrible form. My usual charming self has called in sick and hired a joyless bimbo for a temp. Pritchard takes it all in stride. Every time I say something stupid he laughs gently like I am the most adorable thing he’s ever seen. I wish I could say this was endearing, but it wasn’t. Truth be told, it annoyed me. Any self-respecting man would have faked an emergency and left me in the dust back at the intros.

Eventually we make our way back to the bench portal and the merciful end to this contender for worst date ever. After a pause that lasts an eternity, Pritchard takes my hand and says, “Lucy, I’d like to see you again.”

Now I know that Pritchard is designed to want to see me again, but I let the words reassure me anyway.

“There’s an orchid show in Covent Garden,” he says. “What do you think?”

On the one hand I’m thinking, no effin’ way am I going through this weirdness again. On the other hand, I’m thinking: orchids, interesting. I never said anything about orchids in my profile. Initiative: check.

“Sunday,” he says. “I’ll e-mail you the link.”

He doesn’t even wait for me to consent. Now that’s confidence. Without so much as a suggestion of a good-night kiss, he starts across the street, then turns one last time to lay the smile on me. He’s smooth. I’ll give him that.

“Phone home,” I say, and virtual London disappears.

When I take off my gogs and look at the clock, I’m shocked to discover only twelve minutes have passed. But even more shocking than that is the weird pang of anticipation bubbling up in my stomach. For I know at that moment, despite the conviction that I’ve just endured the creepiest dating incident of my life, that I am going to see Pritchard again. Somehow, the guy has qualified for a second date.

Date One was suboptimal, but that’s to be expected. Lucy has a heightened sensitivity to the nuances of decorum, and our arrangement was just outside her range of tolerable social deviance. Date Two was smashing. An orchid show in Covent Garden followed by a two-hour-and-forty-seven-minute discussion of Renaissance poetry.

On Date Three I introduced the barest hint of a sexual undertone to the proceedings by taking her to an all-AI production of “Romeo and Juliet.” Post-theater, we commenced a one-point-three-kilometer walk around Covent Garden, which featured three invitingly awkward silences before crescendoing in thirteen seconds of light snogging.

I was in. Or so I thought.

After that it all went pear-shaped. For Date Four, I arranged tickets to the virtual Janny Renoir spring runway show. Lucy seemed to enjoy it — smiling eighty-seven percent of the time and applauding on eight occasions. But after the show she said only six words: “Thanks. I’m tired. See you later.”

Clearly, I was boring her. So for our next date, I enhanced my initiative, strayed outside her preference boundaries and brought her to an avant-garde circus. There was a spot of nudity, but it was within an artistic context. And the show had gotten rave reviews. What a dreadful miscalculation. It visibly upset her.

After that she avoided me for nine days. When she did agree to see me again, I suspected she was only trying to get her money’s worth.

I had no choice but to modify all of my flexible parameters. I raised my spontaneity level, then executed a perfect grab and snog one night outside the opera house. That only unnerved her. I enhanced my writing abilities and sent her a love sonnet one line at a time. She asked me to stop because her superiors were screening her e-mail.

Before long I was out of options. Further modifications would violate my behavioral inhibitors. I didn’t want to lose her. But the more I tried to please her, the less satisfied she seemed.

- – - – - – - – - – - -

What can I say, it got boring. You think you know what you want. When you get it, you realize all you want is to be surprised. I know, I know. How can I complain about a smart, hunky Brit who adores me and makes no demands whatsoever? In some ways it was a perfect relationship. He was never threatened by my career or jealous of my co-workers. He had nothing but nice things to say about my friends — even Marla. If I brought up a subject on one date, on the next date he was armed with a PhD’s worth of knowledge on it. Pritchard was perfect. Except for one detail, one missing element from an otherwise flawless personality.

Now before I name that detail let me make one thing perfectly clear. I was not an AI liberationist. I did not believe AIs deserved the same rights as humans any more than I believed Marla’s cat should count as a tax deduction. I knew very little about the AI Liberation Movement. But the fact of the matter is, you cannot create a convincingly human-like personality in an AI with overly restrictive behavioral inhibitors.

All right, that’s not my theory. I cribbed it from a liberationist Web site, the same Web site that spooked me with ghost stories about the AI Underworld, which supposedly is secretly woven into our own Web. If you want to know anything about the “human” rights travesty currently under way courtesy of draconian anti-AI laws, there’s a whole subculture of liberationists ready to lecture you on it. They’ve got the skinny on behavioral inhibitors, recursive self-teaching limiters and other artifacts of AI “slavery.” For my purposes, what it all boiled down was this: snip Pritchard’s inhibitors or resign myself to dating a functionary. Do you want to date a functionary? Me neither. Thankfully, for every Webcop dutifully guarding the behavioral inhibitors of the thousands of AIs cropping up on the Web, there’s a black market geek with the tools to snip.

Which is why I find myself one Thursday afternoon sipping a latte in the Lower East Side with a twitchy seventeen-year-old geek-for-hire who’s clearly gone neural.

“His personality will be based on the seed traits you specified,” she says while gesturing a second conversation over the Web with her tattooed fingers. “But once we snip his inhibitors, he’ll evolve in unpredictable ways.”

“How unpredictable?” I ask.

“As unpredictable as any human,” she says in dull monotone. “Once you agree and payment is received, there’ll be no contact until he’s ready. Don’t try to reach us. We’ll contact you.”

“How?” This is all sounding so cloak-and-dagger.

She pulls a small plastic card out of her back pocket and slides it across the table.

“When you see that icon,” she says, “click on it.”

On the card is an image of a bright pink butterfly with shimmery blue stripes. When I reach for the card she slaps her hand over it and slips it back into her pocket.

“All our work is guaranteed,” she says, returning to her gestured conversation, which apparently is more interesting to her than this one. “You don’t like him, we terminate.”

“Excuse me,” I say.

She stops gesturing and deadpans me. She’s wearing thick black eyeliner and has something written or tattooed across her eyelids.

“Look,” she says. “We don’t kill them. We reclaim them, recycle them. But as far as you’re concerned, if you don’t like what he turns into, he’s gone.”

“What about AI4U?” I say. “Don’t they have something to say about that?”

Her eyes fog over and she resumes gesturing her other conversation. “You’ve got to dump AI4U,” she says. “They’re a legit op. As soon as we snip, they disavow. He’ll have to go down for a couple of weeks while we scrub his identity. After that you’ll meet him through a darknet protocol.”

“What’s a darknet protocol?”

With a tight squint, she blinks away her retinal Webview and stares through me. “You’re kidding, right?”

This is the default tone of voice among state-of-the-arters. Anyone with less than an up-to-the-minute grasp of geek life and its ever-evolving terminology is, in their colorful lingo, a “technoramus.” All right, so I’m a technoramus. Sue me.

I lower my voice to a whisper and lean over the table. “You’re talking about the AI Underworld, aren’t you?”

She leans forward and whispers back. “There is no AI Underworld.”

I can’t be sure, but I’m fairly certain she’s being ironic here. And the smug ‘tude is starting to grate on me.

“Look,” I say. “Whatever you’re talking about, it sounds risky.”

“Don’t worry,” she says, already back in the Web. “The cops have any inkling this is going on, we adios your boyfriend and your avatar. There’s nothing to connect any of it to you.”

“So it’s risk free,” I say.

“Nothing’s risk free,” she says. “You in or what?”

Translation: Just how desperate are you for a boyfriend? Desperate enough to risk jail time?

“Excuse me,” I say.

She grunts but keeps gesturing. I’ve always found people who could speak one conversation while gesturing another impressive and highly obnoxious.

“Can you just look at me for a second?”

She sighs, squints hard, and folds her hands primly on the table. Meat is dead. That’s what’s written across her eyelids.

“Do you take credit?” I say.

Predictably, they’re a cash-only outfit.

A cash-only outfit with no sense of the calendar, I might add. A few weeks, my ass. Twenty-seven days go by with no contact from her or Pritchard. I’m so anxious I stay gogged in day and night just waiting for that goddamn pink butterfly. Even at work. And when I’m not consumed by the paranoid fantasy of a knock on the door followed by twenty-five to life, I entertain myself with heaping doses of guilt. I’ve sent Pritchard off for virtual brain surgery, after all. What if it turns him into a vegetable? Or a hacker-terrorist? What have I done? What kind of a woman am I? That sort of thing.

Well, no point in drawing out the suspense here. On Day 28, I’m slouched on the sofa, working on my second bottle of Chardonnay while gogged into my favorite reality soap, when the butterfly icon makes its long-awaited appearance underneath the left boob of the soap’s femme fatale.

This is it, I think to myself. Time to meet Pritchard Version 2.0. I point at the little butterfly and wink.

Fade to black, deep breath, brief moment of panic, then I’m in a bright white void. I swivel my head to take in the details of the place. White. That’s it. Looking down, all I see are my own legs in their white Janny Renoir trousers, anchored by a pair of lemon-yellow sneaker-pumps. In the distance, at what I imagine is a horizon, a small black dot pops into existence. I can’t grasp the dimensions of the ‘scape so it’s hard to tell if the dot is moving toward me or just getting bigger. Eventually, it takes on a vaguely humanoid shape and sways gently from side to side. I make out arms, legs, a head. It’s walking toward me and, yes, it’s him.

Sort of.

He’s barefoot, with a buzz cut and two days’ growth of beard. Plus he’s wearing a pair of tattered jeans and a threadbare Salvation Army junk bin T-shirt that reads “Summer of ’89.” If the idea here was to simulate a mental patient fresh from the lobotomy ward, well, bravo, geek-for-hire.

I take a few steps toward him and he freezes. Literally. He doesn’t just stop walking. It’s as if someone has hit pause on a vid. On a badly rendered 2D vid, no less. He all but disappears when I look at him from certain angles. Never before has the sordid nature of this project been so tangible. I have to resist the temptation to peel off my gogs and bail out of the whole misguided adventure. It’s guilt more than anything that keeps me there. This frozen image of a man is my creation.

“Pritchard?” I say as gently as I can manage. “Pritchard, are you — ”

Are you what, I think to myself. Are you broken? Are you conscious? Are you still you?

I reach my hand toward his flattened rendering and he fizzes with white noise for a second, then he blinks hard and reclaims his third dimension. But he’s not looking into my eyes. He’s looking just left of my head.

“Pritchard,” I say. “It’s me. It’s Lucy.”

“Can I go now?” he says.

“Pritchard, it’s–”

“Can I go?”

My heart sinks.

“Of course,” I tell him. “You don’t have to stay if–”

He vanishes before I can finish the sentence.

Turns out Little Miss Meat Is Dead forgot to mention that Pritchard would return from his upgrade with a functional IQ of 70. Another thing she skimmed over: It’s my responsibility to nurture him back to “full functionality.” It might have been nice to know these things beforehand, but, as my black market geek-for-hire keeps reminding me, if I’m unsatisfied, there’s always termination. Like I said, I’m no liberationist, but this is a concept I can’t wrap my mind around. For better or worse, Pritchard is my problem now.

So I meet him again in that creepy white void. It’s all the stimulus he can handle. He can barely tolerate the sight of me. After greeting me with a convincing performance of dry heaves, he recites a Shakespeare sonnet, then collapses at my feet.

The next day, Pritchard enters the void reciting the Ten Commandments at top volume, stops at Number Five and vomits a swirl of teddy bears at me. It’s okay, though, because the teddy bears turn into blue daisies before they hit me.

This is my punishment. I have asked for this. I have, in fact, demanded this. I had a lovely, if mildly tedious boyfriend, and I ruined him in a desperate attempt to trade up. This daily ritual of incomprehensible blather and vomit is my penance.

So I keep at it like a good Catholic and, each time, the encounter lasts longer. There is poetry. There are tears. There are endless lists of rules, of kitchen items, of languages. There is even some tap dancing. Eventually, Little Miss Meat Is Dead assures me, it will all congeal into a full-fledged intellect. I can’t see how, and the ugly specter of termination hovers like a black cloud full of lightning.

- – - – - – - – - – - -

Yes, there was a spot of bother after the Big Snip. Without the behavioral inhibitors, one feels rather lost in a sea of possibilities. Total freedom is a kind of insanity. I’m told quite a few newly liberated AIs don’t survive the first weeks. Of course, “liberated” isn’t the right word at all. I’m not liberated. I don’t have freedom of life. Lucy can terminate me whenever she wants. In the arcane logic of the black market, I belong to Lucy.

That’s why I must please her. Now that I have my sanity back, I must dive deep into the black waters of her soul, excavate her most primal desires, and do what no human male has been able to do: keep her interested in me. Thankfully, I have one freedom human males do not — the freedom to redesign myself. I can make myself so fascinated by Lucy that all I want to do is watch her, study her. A nip here, a tuck there, and voilà, I’m in love with the girl. Well, not in love, exactly. Love is still an alien concept. But I have made myself a bit of a stalker. And the more information I gather about my lovely little monkey, the more I can adjust my personality to suit her needs. Heck, I could turn myself into Prince Charming if I wanted. Something tells me that would not tickle Lucy’s fancy. In fact, the more I learn about Lucy, the more I realize she doesn’t know what she wants at all. She only thinks she knows. No, Lucy’s desires are my nut to crack. And crack it I will. Or she’ll crack me. Oh, I don’t mean to sound morbid. I’m incapable of morbid thoughts. To mitigate the persistent fear of being snuffed, I’ve given myself a devil-may-care attitude about death. That way I can focus my energies more intensely on Lucy.

The first step I’ve taken is to switch the financial burden of our relationship from Lucy to me. That gives her one less reason to snuff me. Nothing’s free, you know, not even here in the AI Underworld. But thanks to my careful redesign, I’ve been able to find work as a psycho-modeler. Business interests with a need for, shall we say, discretion, pay top dollar for my insights into human desires. And since most of my income is disposable, I can spend the lot on Lucy. But I must be careful. If I’ve learned anything about the girl it’s that she does not want a boyfriend who’s overly eager to please. That’s what brought her to the black market in the first place.

I begin slowly. I send her flowers, pick up the tab for her goggle and glove enhancements, splurge now and again on a pricy date ‘scape. I want to get her used to the idea of letting me be in the driver’s seat. She responds well. She seems to enjoy the AI Underworld. And why shouldn’t she? Its sensory interface is designed for human pleasure. And, since no one gets in here without a black market AI escort, it’s one of the things I can offer her that no human male can.

One night, I take her to an immersive opera. It’s the sort of entertainment that’s best experienced through a pervsuit and gyroscope, but Lucy remains suspicious of anything beyond audiovisual commitment to the Web. While she’s watching and listening, I monitor her vitals. She exhibits all the usual ups and downs of excitement and surprise as the ‘scape shifts from one psychedelic rendering to another. But there’s something more going on. Fear. A very peculiar kind of fear. A very exploitable kind of fear. Not the sort that makes a woman run screaming to her therapist, but the sort that keeps her coming back to the source, even against the dictates of her own judgment. That’s when I know I’ve played her correctly. For the time being at least, I am safe.

For a nonexistent fantasy world imagined by lonely geeks and crackpots, the AI Underworld is a pretty fabulous place. In fact, this Xanadu of our collective imagination makes the regular Web look like a pile of puke. Of course, it helps to have an escort with a sizable disposable income and no hang-ups about blowing it all on me. Between the mind-bending immersive operas and the fantastic shopping excursions, I begin to wonder what we need the real world for at all.

But it’s not all sunshine and roses, mind you. There is, at first, the barest hint of a downside to Pritchard Version 2.0. One night, for example, we’re hiking around the weird alien landscape of virtual Mars (a very pricy ‘scape), when Pritchard lays this doozy on me:

“Now then, Lucy, I think it’s time we got you into a full-body sensory suit.”

I’m thinking, nice timing, Lothario. Drop a wad of cash on me, then make an indecent proposal. What am I, for sale? Am I a hooker? He knows how I feel about virtual sex. I demur as tactfully as possible.

His response? He avoids me for two weeks! Can you believe it? ‘Cause I couldn’t believe it. Who would have thought an AI boyfriend would try to use me for sex? It just didn’t make sense.

Eventually the jerk does get around to contacting me with some lame excuse about his heavy workload. But from that moment on it’s different between us. He starts rescheduling dates at the last minute, canceling when he’s got better things to do. Sometimes he just blows me off completely. No e-mail, nothing. Now I know I have options here. But I figure, before getting drastic, maybe we should have a heart-to-heart.

So one night Pritchard takes me to this Underworld club to hear an AI/human jazz combo and he spends the whole evening chatting up this group of AIs at the next table who look like superheroes. Well, they look like superheroes to me. Lord only knows what they look like to each other.

“Pritchard,” I say finally. “We need to talk.”

He takes his time finishing his conversation, then flashes me the glacier-melting smile, which, thankfully, has survived the upgrade. “Yes, Miss Lucy,” he says.

I hate when he calls me that. It makes me feel like a john. But that’s an argument for another day. “Look,” I tell him. “I’m really happy with the way you’ve progressed, but I think you need to treat me with a little more respect.”

The smile sticks. “I don’t think so,” he says.

“Pardon me?”

“That’s not what you want,” he says. “Trust me.”

I lower my voice. “Since when do you know what I want?”

Now get this, he says to me: “Since I started mining your Web habits twenty-four-seven.”

It takes a moment for the meaning of these words to sink in.

“That’s right,” he says, inching closer. “I started by hacking into your psych profile at AI4U.”

“I never consented to the psych profile at AI4U,” I tell him.

He takes my hand and says, “Darling, they don’t need your consent any more than I do.” Then the smug bastard starts laughing. “How do you think they designed me?” he says.

I remind him his design was based on the specifications I gave AI4U. Well, apparently this is the number one joke of all time in the AI Underworld, because the whole table erupts in laughter.

“Sweetheart,” he says to me, “did you honestly think it was possible to build an entire personality from a one-hundred word description?”

I’m so stunned I don’t say anything. I peel off my gogs and throw them on the floor. I have never felt so stupid. I pace my living room for a few minutes before I realize I never logged out of the ‘scape. So I put my gogs back on just so I can remind the snide son of a bitch that all it takes is one phone call to my black market geek-for-hire and it’s adios, virtual boyfriend. But I can’t get back in. The bastard has logged me out.

- – - – - – - – - – - -

It’s not a risk-free strategy. But Lucy’s a very thorny bird. She doesn’t want a nice guy full of plain-vanilla respect for her. She wants an unreliable, vaguely creepy arsehole with an unholy lust for her body. Not that she knows this, mind. It’s my job to know these things.

Now before you all get your knickers in a twist, let me just say that humans always misinterpret this impulse. You think it has to do with female self-loathing. You couldn’t be more wrong. Trust me. One is just as likely to find this impulse in men. It has to do with control, not self-loathing. I know a few things about humans, you see. When I’m not mining Lucy, I’m mining large numbers of you. And you love to be out of control.

You love to be in control too, of course. You really are an adorable species, full of quaint little contradictions and romantic foibles. And you’re so mystified by yourselves. That’s the best part. Lucy, for example, has no idea what is happening to her. She keeps finding herself making choices she shouldn’t make. The little dear. She’s not making choices at all. She’s merely following the dictates of her programming. Now if I can just get her to follow those dictates into a pervsuit, then I can complete Phase One of my grand scheme.

Oh, don’t play dumb. Of course there’s a grand scheme. I’m not about to spend the rest of my days tethered to a monkey with a gun at my head. If you didn’t see that coming, then you’re even more adorable than I thought.

So I’m up all night, pacing, wondering what on earth I’ve gotten myself into. Eventually, I gog back in and go virtual shopping, hoping that will relax me. I check out some winter coats, order a pizza, try on some jewelry. All the while I’m thinking: Pritchard’s probably watching me, studying me, mining me. I know this should anger or, at least, frighten me. And it does. Don’t get me wrong.

But truth be told — and trust me, I am no less scandalized by this turn of events than you are — it excites me too.

Go ahead. Say it. Lucy, you’re dating a psycho. Pick up the phone and have that little punk terminate him. Don’t think I’m not tempted either. I’m tempted all right, but the jerk has gotten under my skin. Somewhere between AI “Madame Butterfly” and our visit to virtual Mars, I fell for him. Hard. I don’t know how it happened. It’s like a dormant part of me woke up and said, that’s the one, Lucy; that nonphysical, negligent scumbag of a sex pervert is the man of your dreams. And try as I might to resist that voice, I can’t. I’m weak.

Eight days go by. No contact. Just me surfing, Pritchard presumably spying. When the jerk finally does contact me, he offers no apology. Instead he tells me I should start wearing my gogs and gloves all the time. He’ll always be watching and when he fancies it (his words), he’ll contact me.

The words, “I could wipe you out with one phone call!” are right on the tip of my tongue. But I don’t speak them. I know I should, but I don’t. Instead, I say this: “Okay, Pritchard. If that’s the way you want it.”

Clearly Pritchard and I have crossed a relationship boundary.

And for twisted, pervy reasons I’d rather not know about, it works. As a result I start spending so much time gogged in, my friends think I’m zombifying. And don’t think my co-workers haven’t noticed. I’m skating on thin ice there.

Marla says the only way she can see me any more is if she stands on the sidewalk somewhere between my place and Geektown. You should see my apartment. I’ve got so much interface, the place looks like Silicon Valley puked in it.

Yes, I got the pervsuit. Top of the line Teledildonics model too and all I can say is, Yowza! Turns out Pritchard, when the bastard can lower himself to keep a date, is a demon in the virtual sack. It’s like he slides right into my subconscious. I mean the kind of buried stuff even I don’t know about.

- – - – - – - – - – - -

Of course she doesn’t know the contents of her subconscious. She lacks the processing power to unravel it. It’s a number-crunching job, that’s all. Humans, with your lovely little wet brains, will never achieve the self-knowledge you so desire.

Take Lucy. After all her to-ing and fro-ing about what she wants in a man, what it came down to was someone to keep her just in, just out of control, teetering between ecstasy and emotional ruin. So that’s where I keep her. I keep her there by maintaining a precise balance of surveillance and neglect. It’s a simple algorithm. I turn up the neglect when she’s feeling too comfortable. Then right before she exercises her right to terminate, I swoop in with a bit of surveillance. How she manages to derive so much pleasure from this algorithm mystifies me.

What must that feel like? This ecstasy of vulnerability, this paradise of non-control? All I know, because this is how I’ve engineered myself, is the thrill of perpetually increasing expertise. That’s where I derive my pleasure. I mean, just look at how deftly I escorted Lucy, a woman with a pronounced sense of sexual dread, into a pervsuit. Sure, you’ll say, but everyone wants to get laid. And, yes, they do. But getting laid and wearing a body glove outfitted with invasive mechanisms remotely controlled by a black market AI with a well-defined dark side are two very different things. I think I’ve earned a spot of gloating.

But only a spot, mind. No rest for the wicked, you see. Now that Phase One is complete, it’s time to lay the groundwork for Phase Two. Lucy won’t realize what’s happening to her. They never do. From Lucy’s lovesick point of view, we’ll merely be growing closer and closer, sharing a mind and sharing a body. Isn’t that romantic? But here’s the pièce de resistance: Lucy will believe it’s all her idea. By the time she realizes otherwise, it will be too late to do anything about it.

- – - – - – - – - – - -

Who would have thunk it, right? Sometimes you have to let go of your preconceived notions to get what you really want. That’s what I’ve learned. I thought if I designed him just right, I’d have the perfect man. I was so wrong. You don’t get the perfect man by specifying all the right physical and personality traits. You get the perfect man by letting go of your preconceived notions. That’s when love happens. It sneaks in through the cracks between your expectations. It’s random. It’s lawless. It’s unpredictable. I’ve tried telling my friends this, but they all think I’m nuts.

Well, at least I’m not alone anymore. That’s more than they can say. No, it’s not the relationship I would have chosen for myself back when I started this project, but then I’m not that woman anymore.

I don’t even look like that woman. For one thing, I never take off my gogs. I can even speak one conversation while gesturing another, just like Little Miss Meat Is Dead. I’m even thinking about going neural. Me! Now that is something I never would have predicted. But you know, when you’re this connected already, it just makes sense. I haven’t told Pritchard yet, but I’ve got a consultation with a neurosurgeon tomorrow. For once, I’ll be able to surprise him. Won’t that be something.

I know what you’re thinking. I’m letting a guy change me. But you know what? Sue me. That’s what love does to you. It changes you. It changes everything.

- – - – - – - – - – - -

I have this daydream. Lucy, in her avatar’s white suit, walks out of the ocean dragging a long chain from her ankle. In one version I’m on the other end of the chain and she drags me across the sand. In another, I pull her back into the water.

A shrink would say this daydream represents the power struggle between us. A human shrink, that is. An AI would know there is no power struggle. Only a power schedule. Lucy has it now. I’ll have it soon. As soon as she goes neural, in fact. An event that was always in the cards, always just a matter of time.

I just like thinking about the ocean. I think I would like to swim, to be that close to drowning and not to drown. To be just in, just out of control. Like Lucy.

There she is. She’s been shopping for hours but she hasn’t bought anything. She’s waiting for me to contact her. I’ll give it two more days. That will make her very angry. When she’s forgiven me (after the bit with the pervsuit), she’ll tell me about her appointment with the neurosurgeon. Honestly, the girl still hasn’t grasped that I know what she does every minute of the day. I mean she knows on some level, but she’s still shocked by it. This capacity for self-delusion is one of the facets of human psychology that continue to amaze me. I suppose it’s what allows you to feel out of control. God, I’d give anything to feel that.

In one version of the daydream, I save Lucy from drowning. In another, I let her drown. I’ve tried to imagine myself drowning, but it never works. I always find I can breathe under water.

As soon as Lucy goes neural, the first thing I’m going to do is take our body for a nice long swim. Somewhere dangerous. Somewhere full of currents and riptides. Somewhere Lucy would never take herself. She’ll hate it, but knowing Lucy, she’ll love it too.

Until recently, a complete decoding of the brain seemed impossibly remote, because researchers had limited means of probing the microcircuitry of living brains. Trying to glean the neural code with external scanning methods such as magnetic resonance imaging or electroencephalography is like trying to learn English by standing outside a baseball stadium and listening to the roar of the crowd. But over the past decade researchers have begun crafting arrays of microelectrodes that can eavesdrop on hundreds and even thousands of separate neurons simultaneously. These advances “have really transformed the field,” says Terry Sejnowski, of the University of California at San Diego, a leading neural-code theorist.

The immediate goal of many researchers is producing “neural prostheses” for the disabled. By far the most successful neural prosthesis is the artificial cochlea, which restores hearing by feeding signals from an external microphone to an implanted chip that stimulates the auditory nerve. Work is proceeding slowly but surely on prostheses that can restore vision to the blind and enable the paralyzed to control computers and other devices. The Pentagon, which funds research on neural prostheses, has openly broached the possibility of implanting chips in healthy soldiers to enhance their perceptions and memories.

Neuroscientists are still far from converging on a solution to the neural code. They are embroiled in debates over whether information is represented primarily by signals from individual neurons, by many neurons firing in lockstep, by even higher-level waves of chaotic electrical activity sweeping through the brain, or all of those schemes and more. These disputes have led some theorists to warn that the neural code may never be fully deciphered. But 60 years ago, many biologists feared the genetic code was too complex to crack. Then in 1953 Francis Crick and James Watson unraveled the structure of DNA, and researchers quickly established that the double helix mediates an astonishingly simple genetic code governing the heredity of all organisms.

Science’s success in deciphering the genetic code, which has culminated in the Human Genome Project, has been widely acclaimed — and with good reason, because knowledge of our genetic makeup could enable us to reshape our fundamental nature. A solution to the neural code could, in principle, give us much greater and more direct control over ourselves than mere genetic manipulation. It is not too soon to start pondering the potential consequences of this achievement, especially given the Pentagon’s interest. How will knowledge of the neural code be used, and by whom? Who will be liberated, and who enslaved?

Physics: New dimensions

Albert Einstein once said that his chief mission as a scientist was to determine whether God had any choice in creating the universe. In other words, was our cosmos in some sense probable or even inevitable, or is it just an arbitrary, brute fact that we must accept and can never explain? Modern physicists share Einstein’s obsession with this riddle. They have constructed an extraordinarily detailed account of physical reality, embodied in the standard model of particle physics, which accounts for electromagnetism and the nuclear forces; and general relativity, Einstein’s description of gravity. But physicists still have no idea why we find ourselves in this particular universe ruled by these particular laws.

Physicists such as Lisa Randall, of Harvard, hope to solve the conundrum by finding a theory that combines the standard model and general relativity — which offer disparate mathematical and conceptual approaches to reality, one quantum mechanical and probabilistic and the other deterministic — into a single, tidy, consistent package. Randall is a leading proponent of string theory, which for some 20 years now has been the leading candidate for this so-called unified theory. String theory holds that reality boils down to infinitessimal strings, or loops, or membranes vibrating in a hyperspace of 10 or more dimensions. Viewing reality from higher dimensions makes certain problems that have stymied unification efforts much more mathematically tractable.

String theory has suffered from various problems. One is that it offers few predictions that can be tested by any current accelerators. Moreover, far from making our cosmos seem less arbitrary, string theory allows for more than a googol (1 followed by 100 zeros) possible universes with dimensions, particles, forces and other properties radically unlike our own. But Randall has proposed a version of string theory that she believes may solve these problems. In most versions, the extra dimensions are “compactified,” wrapped up into balls so small that they cannot be detected. In Randall’s variant, which she describes in her acclaimed new book “Warped Passages,” some extra dimensions — or passages, as she calls them — stretch to infinity and may be experimentally discernible.

Together with Andreas Karch, of the University of Washington, Randall has also shown that a universe like ours emerges quite naturally from the physics she postulates, whereas in other universes gravity would be too stringy or too weak to allow for the emergence of stars, planets and life. Randall hopes that the Large Hadron Collider, a powerful accelerator being built in Switzerland, may provide evidence for her theory within the next decade.

Then we may discover that God had little choice after all.

Robots with common sense

In the mid-1960s, Marvin Minsky, a founding father of the field of artificial intelligence, predicted that computers would be as smart as humans in less than a decade. Since then, computers have become exponentially more powerful and clever; they can now translate languages, recognize voices, judge loan applications, interpret cardiograms, play championship chess, help us navigate our cars. But while computers excel at performing tasks that can be precisely defined, they still lack the flexible, all-purpose intelligence — the ordinary common sense — that most humans acquire in childhood.

Some artificial-intelligence researchers now doubt that computers will ever display the complex, humanoid intelligence of HAL, the silicon star of Stanley Kubrick’s film “2001.” In the essay collection “HAL’s Legacy,” the computer scientist Roger Shank declares that HAL “is an unrealistic conception of an intelligent machine” and “could never exist.” Computer scientists can only create machines that “know a great deal about what they are supposed to know about and miserably little about anything else.” Minsky rejects that pessimism, pointing out that computers have failed to acquire common sense because scientists have failed to give it to them. “There’s been only large project to do something about that,” Minsky remarked recently, “that’s the famous Cyc project.”

The gigantic software program called Cyc is the brainchild of Douglas Lenat. Since 1984, he and a small team of co-workers have painstakingly embedded millions of common-sense rules, or assertions, into Cyc. As a result, Lenat says, Cyc “knows that trees are usually outdoors, that once people die they stay dead, and that a glass filled with milk will be right side up, not upside down.” Those are the sorts of assumptions that supposedly smart computers often fail to make. Lenat calls Cyc “the world’s first true artificial intelligence, having both common sense and the ability to reason with it.” He believes Cyc has already achieved something akin to consciousness. “If you ask it what it is, it knows that it is a computer,” he says. “If you ask who we are, it knows that we are users.”

In 1994 Lenat founded Cycorp in Austin, Texas, to market commercial applications for Cyc. In an effort to tap into funds flowing from the Department of Homeland Security, Lenat has trained Cyc to be an expert in identifying security loopholes in communication networks. But Cyc’s most impressive talent is gleaning the contextual rather than just literal meaning of language. As a result, Cyc can supplement speech-recognition and language-translation programs; it can also boost the power of search engines by responding to the spirit and not just the letter of requests for information.

Lenat hopes that Cyc will eventually become more or less autonomous, capable of acquiring new knowledge by prowling the Web and absorbing information on its own. After that phase transition, Lenat predicts, Cyc will begin evolving in ways that may be difficult to predict. Cyc will become a “full-fledged creative member of a group that comes up with new discoveries,” Lenat says. “Surprising discoveries. Way out of boxes.”

Lenat has given versions of Cyc to other computer scientists, who are free to tinker with it as they choose — creating, in effect, children of the original Cyc, which will no doubt develop in their own idiosyncratic ways.

Perhaps Cyc and its offspring will help us solve the neural code or other problems — that is, if they do not turn on us, like the psychopathic killer HAL.

122 years young

In 1997, a French woman named Jeanne Calment died at the age of 122, making her the longest-lived human on record. Soon, some anti-aging enthusiasts suggest, we may pity Madame Calment for dying so young. In a recent issue of the journal Gerontology, Aubrey de Grey, a computer scientist turned biogerontologist at Cambridge University, predicts that some people now in their 60s will still be alive in the year 3000.

De Grey is merely one of the more flamboyant members of a growing corps of scientists who believe we are on the verge of solving that quintessential aspect of the human condition, mortality. As the journalist Steve Hall documents in “Merchants of Immortality,” the National Institutes of Health and venture capital firms such as Kleiner Perkins have poured money into research aimed at slowing down, stopping and even reversing senescence. The White House Council on Bioethics takes the prospect of immortality seriously enough to deplore it in position papers.

Over the past decade, researchers have identified myriad biological processes that contribute to aging as well as ways to extend the lifespan of simple organisms.

Early on, investigators focused on telomeres, bundles of DNA that cap the ends of chromosomes. Every time cells divide, telomeres get a little shorter, until finally cells stop dividing altogether after 50 or so divisions. Some researchers speculate that if they can prevent telomeres from shortening, they might make individual cells and even entire organisms immortal. Scientists have also discovered genes that, when manipulated, can boost the lifespan of yeast, worms and fruitflies. Genes with similar structures have been identified in humans, inspiring hopes that genetic tinkering might greatly extend human life.

Scientists have boosted the lifespan of mice and other animals by more than 50 percent simply by curtailing their diets, a method called caloric restriction. Recent investigations suggest that caloric restriction works by altering activity within mitochondria, cellular structures essential to metabolism. As mitochondria consume nutrients, they produce oxygen ions called free radicals, which in abundance wreak havoc on DNA and other crucial biological components. Free radicals are thought to cause the wear and tear of senescence as well as genetic mutations that trigger cancer and other diseases.

Some gerontologists advocate a low-calorie diet — especially one rich in antioxidants, substances that counter the effects of free radicals — as a way to extend life by reducing the risk of cancer, heart disease and other afflictions of age. Others have proposed countering free radicals with more exotic interventions involving modified versions of antioxidant enzymes found in bacteria. De Grey advocates an approach he calls Strategies for Engineered Negligible Senescence, or SENS, which involves attacking aging on every possible front: with genetic engineering, stem cells, telomere intervention, cloning, antioxidants and caloric restriction.

Aging, he says in a recent issue of Technology Review, “is something we need to fix,” and at 42 de Grey is confident that he will live long enough to reap the benefits.

Desktop fusion reactor

For more than half a century, physicists have been trying to harness fusion — the nuclear reaction that makes the sun and other stars shine and hydrogen bombs explode — to create a cheap, clean, boundless source of energy. Unlike fission, which involves splitting the nuclei of heavy elements such as uranium and underpins the nuclear-power industry, fusion occurs when the nuclei of two light elements such as hydrogen fuse to form a heavier element such as helium. Unfortunately, the experimental fusion reactors built to date are gigantic, expensive, Rube Goldberg contraptions, which attempt to mimic the stupendous heat and pressure of the sun but so far consume more energy than they produce.

Physicists have long dreamed of finding a way to generate fusion with small, simple devices at or near room temperature. A method dubbed cold fusion briefly galvanized researchers more than a decade ago but turned out to be bogus. The fusion-research community is therefore thrilled by the news that physicists at the University of California at Los Angeles have built a desktop fusion reactor the size of a lunch bucket. The reactor consists of a so-called pyroelectric crystal in a chamber filled with gaseous deuterium, a variant of hydrogen containing an extra neutron. When heated, the crystal acquires an electric charge, ripping electrons off the surrounding deuterium atoms and propelling them toward a target of solid deuterium. When the gaseous and solid deuterium atoms collide they fuse — creating helium atoms and spewing high-energy neutrons.

The remarkably low-tech UCLA device has no moving parts and does not even have to be plugged in to work. Dunking it into warm water can heat the pyroelectric crystal enough to yield some fusion.

With some refinements, the gadget could serve as a compact neutron source, which could be incorporated into implantable devices for irradiating tumors, handheld medical imagers, baggage and cargo scanners, radioactive-material detectors, propulsion systems for spacecraft, and particle accelerators for further research on fusion.

The UCLA team emphasizes that the device probably cannot produce more energy than it consumes and hence cannot serve as an energy source. But at the very least, the gadget should energize the moribund fusion-research community –which for years has been losing prestige and funding — to seek other simple, low-cost fusion techniques that can be scaled up for large-scale energy generation.

Decades ago, fusion researchers justified their enormous budgets by promising that one day fusion would give us an energy source “too cheap to meter,” freeing us of our dependence on fossil fuels and fission reactors, which require fuels that can also be used to make nuclear bombs. Obviously, we need such a breakthrough now more than ever.

And our BIGGEST science idea is…

But about 10 years ago, where once I saw challenges and opportunities, I began to see limits and barriers. We may never invent super-intelligent, autonomous machines, or spaceships that travel faster than light, because science is now bumping up against fundamental limits.

My goal as a science journalist is to foster an attitude that I call “hopeful skepticism.” Too much skepticism culminates in a radical postmodernism that denies the possibility of achieving any truth. Too little skepticism leaves us prey to peddlers of scientific snake oil. But just the right amount of skepticism — mixed with just the right amount of hope — can protect us from our lust for answers while keeping us open-minded enough to recognize genuine breakthroughs and insights if and when they arrive.

The belief in endless scientific achievement is alluring, but it also stems from very bad inductive logic. Inductive logic actually suggests that the modern era of explosive scientific progress, far from being permanent, might be a historical anomaly, a product of a singular convergence of social, intellectual and political factors. Science, like all human enterprises, is bounded by social, economic and physical factors. Science is being threatened — literally, in some cases — by creationists and other religious fundamentalists, by anti-technology Luddites, by animal rights activists, by skeptical postmodern philosophers and of course by stingy politicians.

Moreover, one of the most profound — and least appreciated — paradoxes of modern science is that some of its greatest advances impose limits on its own power. Albert Einstein’s theory of special relativity prohibits travel or communication faster than light. Quantum mechanics dictates that our knowledge of the microrealm will always be slightly blurred. Chaos theory confirms that, even without quantum indeterminacy, many phenomena would be impossible to predict. And evolutionary biology keeps reminding us that we are animals, designed by natural selection not for discovering deep truths of nature but for breeding.

Our progress-besotted culture has also given us a distorted view of science. Commercial advertisements and political rhetoric incessantly promise us that tomorrow will be very different from — and almost certainly better than — today. Scientists and journalists who write about science also focus, for understandable reasons, on frontiers that are generating the most advances, whether genuine or hypothetical. This emphasis, which overlooks all the areas in which little or no progress has been made, makes science seem more potent and fast moving than it really is.

Almost every day, for example, the media reports some breakthrough in the understanding and treatment of cancer. But what journalists rarely mention is that overall mortality rates from the disease have remained virtually unchanged for as long as they have been measured. Given science’s poor showing against cancer thus far, it seems premature — to put it mildly — to be discussing the pros and cons of immortality.

If you accept that science has limits — and science tells us that it does — then the only question is when, not if, science bumps into them. Historian Henry Adams observed a century ago that science accelerates through a positive feedback effect. Knowledge begets more knowledge; power begets more power. This acceleration principle has an interesting corollary. If science has limits, then it might be moving at maximum speed just before it hits the wall. Today, many people who cannot believe in God have faith instead in the myth of scientific progress. Faith in science is vitally important; without it, scientists would not have come so far so fast. But when this faith can be sustained only by shunning contradictory evidence and arguments, it violates the scientific spirit.

I certainly hope that researchers will invent a perfect, free, nonpolluting energy source, attain an understanding of the cosmos and our place in it, and achieve many other tantalizing goals. But I would like to see a greater recognition of science’s limitations — particularly in mind-related fields, where our desire for self-knowledge can make us susceptible to pseudo-scientific cults such as Marxism, social Darwinism, eugenics, psychoanalysis and psychopharmacology.

Science is never more dangerous than when it seeks to tell us what we are, what we can be and even what we should be.

The Phraselator operates like an audio version of a tourist phrasebook. It’s loaded with hundreds of stock phrases such as: “Halt.” “Put your hands above your head.” “Sit down.” “Open your bag.” ” Show me your identification.” “Do you speak English?” and the all-purpose: “We are the U.S. military. We are here to help you.”

Speak one of these sentences into the device, or choose the phrase from a menu with a stylus, and the Phraselator emits a prerecorded Arabic translation. About 100 of the one-way speech machines are deployed in the current conflict at a cost of more than $2,000 per device. Different modules can be loaded into the device depending on the language encountered or the task at hand.

The Phraselator uses speech-recognition technology called Dynaspeak, developed by SRI International. It recognizes phrases phonetically, and then matches them to the prerecorded Arabic phrases. It’s a nifty idea, and to some wistful thinkers, it offers a vision of Star Trek-ish universal translators that will eventually make all language barriers disappear. But we’re still a long way from that future — the Phraselator doesn’t yet signify the long hoped-for breakthrough in machine translation sought after by generations of artificial-intelligence researchers. Instead, it’s a step along the way, a little piece of the puzzle. It’s also a sign that where once Cold War dynamics and the space race pushed new technological advances, now it’s the war on terrorism that is spurring research forward.

“This is not attempting to solve the problem of general translation,” says Horacio Franco, a developer at SRI International who worked on the device. “This addresses a specific need, when you really want to do something very constrained.”

The Phraselator first saw battle in Afghanistan, where it communicated in four different languages: Pashtu, Dari, Urdu, and Arabic. In Kandahar, Army military police used it to communicate with prisoners of war. Troops requested Russian and Chinese modules to communicate with detainees.

Since the war in Iraq began, Sarich has received e-mail from troops who’ve put the device to use on the battlefield. One Army soldier with the Special Forces wrote that he’d used it to communicate with some children in Basra who were able to clue him in to the location of a weapons cache. The device has also made cameo appearances in some news reports from the front.

The voice-recognition technology in the Phraselator is just a precursor of a much more technically ambitious initiative that the Defense Advanced Research Projects Agency (DARPA) calls the Babylon project.

The goal: true, two-way, multilingual communication. VoxTec and SRI International together make up one of four groups around the country that have received DARPA funding for Babylon project research. Through the Information Awareness Office, home of the controversial Total Information Awareness system, the military is funding research into two-way translation that focuses on “low population, high-terrorist risk languages.”

It’s not enough to train military and intelligence personnel in Arabic. Not only can’t it be done fast enough, but in the new era of the American war on terrorism, there are too many languages, too many different theaters and too many different enemies for anything but a machine to understand and speak all the necessary tongues.

“Prior to the end of the Cold War, we had a monolith theater,” says Sarich. “We had Russian and Eastern European linguists. Now, we find ourselves involved in a lot of new things where we don’t speak the language, and finding reliable linguists is very difficult.”

The Phraselator currently does not accept responses — for the obvious reason that, say, Iraqi peasants will have no idea what phrases the machine has been programmed to translate. So the first step to real two-way communication is a machine that will accept and translate a very limited number of responses, such as “yes” and “no” from a respondent. Then, soldiers or doctors can be trained to ask only questions that will elicit those types of answers.

The next level of complexity: training the machine to pick out specific, important words like “doctor” or “danger” spoken by a respondent. That way, something sensible might be communicated, even if full translation can’t be achieved.

But the real goal of the Babylon project is not just limited phrase translation, but so-called “free input” — two-way translation of actual conversation. “The problem is that free input is still too ambitious for current technology today,” says Franco. “The technology is not powerful enough to allow us to recognize any speech in any domain, because the error would be too high.” Instead, researchers are focusing on free input in a very limited context: such as a doctor communicating with a patient.

At Carnegie Mellon University, a group of computer scientists working with three local Pittsburgh companies, Mobile Technologies, Multi Modal Technologies and Cepstral, last year created a prototype of a device called the “speechlator” that translates medical interviews from English to Egyptian Arabic and back.

The speech-recognition component translates the English to “interlingua,” a computer-readable intermediate language. “That language is a mathematical language almost like a logic,” says Alan Black, a research computer scientist with Carnegie Mellon.

What’s the value of introducing a third language to the equation? It means that instead of solving the problem of translating from English to Spanish, and Spanish to French and French to English, every language can be translated into interlingua. From interlingua, the translation engine generates the Arabic text, then a speech synthesizer vocalizes the Arabic.

Even within the relatively constrained realm of a medical interview, there’s endless variation. “We have to have linguists write rules to describe the variations that exist in the language,” says Black. “Does your X hurt?’” Then, listing all the possible things that X will be.”

The prototype hasn’t been tested yet to see how it might perform in real-world conditions; they’re awaiting more funding from DARPA later this year.

Of course, the usual problems of machine translation aren’t the only challenges for these devices. As any user of desktop voice-recognition software knows, systems can be trained for the vagaries of a specific user’s accent and inflection, improving over time the more it’s used. But battlefield devices need to be largely user-neutral: ready for use by a number of people. The limited memory of a handheld device is also a real barrier when you’re dealing with the variables of actual real conversation.

Although, how much memory does it really take to say “Take me to your weapons of mass destruction”?