The Most Human Human

by Brian Christian

  If, or when, a computer wins the gold (solid gold, remember) Loebner Prize medal, the Loebner Prize will be discontinued forever. When Garry Kasparov defeated Deep Blue, rather convincingly, in their first encounter in ’96, he and IBM readily agreed to return the next year for a rematch. When Deep Blue beat Kasparov (rather less convincingly, I might add) in ’97, Kasparov proposed another rematch for ’98, but IBM would have none of it. They immediately unplugged Deep Blue, dismantled it, and boxed up the logs they’d promised to make public.1 Do you get the unsettling image, as I do, of the heavyweight challenger who, himself, rings the round-ending bell?

  The implication seems to be that—because technological evolution seems to occur so much faster than biological evolution, years to millennia—once Homo sapiens is overtaken, it won’t be able to catch up. Simply put, the Turing test, once passed, is passed forever. Frankly, I don’t buy it.

  IBM’s odd anxiousness to basically get out of Dodge after the ’97 match suggests a kind of insecurity on their part that I think is very much to the point. The fact is, the human race got to rule the earth—okay, technically, bacteria rule the earth, if you look at biomass, and population, and habitat diversity, but we’ll humor ourselves—the fact is, the human race got to where it is by being the most adaptive, flexible, innovative, and quick-learning species on the planet. We’re not going to take defeat lying down.

  No, I think that, while certainly the first year that computers pass the Turing test will be a historic, epochal one, it does not mark the end of the story. No, I think, indeed, that the next year’s Turing test will truly be the one to watch—the one where we humans, knocked to the proverbial canvas, must pull ourselves up; the one where we learn how to be better friends, artists, teachers, parents, lovers; the one where we come back. More human than ever. I want to be there for that.

  If Victory

  And if not defeat, but further rout upon rout? I turn a last time to Kasparov. “Success is the enemy of future success,” he says. “One of the most dangerous enemies you can face is complacency. I’ve seen—both in myself and my competitors—how satisfaction can lead to a lack of vigilance, then to mistakes and missed opportunities … Winning can convince you everything is fine even if you are on the brink of disaster … In the real world, the moment you believe you are entitled to something is exactly when you are ripe to lose it to someone who is fighting harder.”

  If there’s one thing I think the human race has been guilty of for a long time—since antiquity at least—it’s a kind of complacency, a kind of entitlement. This is why, for instance, I find it oddly invigorating to catch a cold, come down from my high horse of believing myself a member of evolution’s crowning achievement, and get whupped for a couple days by a single-celled organism.

  A loss, and the reality check to follow, might do us a world of good.

  Maybe the Most Human Human award isn’t one that breeds complacency. An “anti-method” doesn’t scale, so it can’t be “phoned in.” And a philosophy of site-specificity means that every new conversation, with every person, in every situation, is a new opportunity to succeed in a unique way—or to fail. Site-specificity doesn’t provide the kinds of laurels one can rest on.

  It doesn’t matter whom you’ve talked to in the past, how much or how little that dialogue sparkled, what kudos or criticism, if any at all, you got for it.

  I walk out of the Brighton Centre, to the bracing sea air for a minute, and into a small, locally owned shoe store looking for a gift to bring back home to my girlfriend; the shopkeeper notices my accent; I tell her I’m from Seattle; she is a grunge fan; I comment on the music playing in the store; she says it’s Florence + the Machine; I tell her I like it and that she would probably like Feist …

  I walk into a tea and scone store called the Mock Turtle and order the British equivalent of coffee and a donut, except it comes with thirteen pieces of silverware and nine pieces of flatware; I am so in England, I think; an old man, probably in his eighties, is shakily eating a pastry the likes of which I’ve never seen; I ask him what it is; “coffee meringue,” he says and remarks on my accent; an hour later he is telling me about World War II, the exponentially increasing racial diversity of Britain, that House of Cards is a pretty accurate depiction of British politics, minus the murders, but that really I should watch Spooks; do you get Spooks on cable, he is asking me …

  I meet my old boss for dinner; and after a couple years of being his research assistant and occasionally co-author, and after a brief thought of becoming one of his Ph.D. students, after a year of our paths not really crossing, we negotiate whether our formerly collegial and hierarchical relationship, now that its context is removed, simply dries up or flourishes into a domain-general friendship; we are ordering appetizers and saying something about Wikipedia, something about Thomas Bayes, something about vegetarian dining …

  Laurels are of no use. If you de-anonymized yourself in the past, great. But that was that. And now, you begin again.

  1. These logs would, three years later, be put on the IBM website, albeit in incomplete form and with so little fanfare that Kasparov himself wouldn’t find out about them until 2005.

  Epilogue: The Unsung Beauty of the Glassware Cabinet

  The Most Room-Like Room: The Cornell Box

  The image-processing world, it turns out, has a close analogue to the Turing test, called “the Cornell box,” which is a small model of a room with one red wall and one green wall (the others are white) and two blocks sitting inside it. Developed by Cornell University graphics researchers in 1984, the box has evolved and become more sophisticated over time, as researchers attempt additional effects (reflection, refraction, and so on). The basic idea is that the researchers set up this room in real life, photograph it, and put the photographs online; graphics teams, naturally, try to get their virtual Cornell box renderings to look as much as possible like the real thing.

  Of course this raises some great questions.

  Graphics teams don’t use the Cornell box as a competitive standard, and there’s an assumption of good faith on their part when they show off their renderings. Obviously, one could simply scan the real photograph and have software output the image, pixel for pixel. As with the Turing test, a static demo won’t do. One needs some degree of “interaction” between the judges and the software—in this case, something like moving some of the internal boxes around, or changing the colors, or making one of the boxes reflective, and so on.

  Second is that if this particular room is meant to stand in for all of visual reality—the way a Turing test is meant to stand in for all of language use—then we might ask certain questions about the room. What kind of light is trickiest? What types of surfaces are the hardest to virtualize? How, that is, do we get the real Cornell box to be a good confederate, the Most Room-Like Room?

  My friend Devon does computer-generated imagery (CGI) for animated feature films. The world of CGI movies is a funny place; it takes its cues from reality, yet its aim is not necessarily realism. (Though, he notes, “your range of what’s believable is wider than reality.”)

  Being a computer graphics person brings with it, as most jobs do, a certain way of looking at and of noticing the world. My own poetry background, for instance, gives me an urge to read things against the grain of the author’s intended meaning. I read a newspaper headline the other day that said, “UK Minister’s Charm Offensive.” This to me was hilarious. Of course they meant “offensive” as a noun, as in the tactical deployment of charm for diplomacy purposes, but I kept reading it as an adjective, as though the minister’s creepy unctuousness had really crossed the line this time. My friends in the police force and the military can’t enter a room without sussing out its entrances and exits; for the one in the fire department, it’s alarms and extinguishers.

  But Devon: What does a computer graphics guy look for?

  “Sharp edges—if you’re looking at, like, anything, any sort of man-made object, if it has sharp edges, l
ike a building, or a table: if all the edges are really sharp, then that’s a pretty good sign. If you look in the corners of a lit room—if the corners aren’t appropriately dark, or too dark … Just like complexity of surfaces and irregularities—any type of irregularity, you know. That’s all totally—if it’s computer generated—it’s really hard to do. You look for the quantity of irregularities and regularities, even textures, and things like that. But that’s all pretty basic stuff. At another level, you have to start thinking about, like, light bouncing off things, you know, like if you have, for instance, a red wall next to a white wall, how much of the red gets onto the white, and that’s the sort of thing that can sort of throw you off.”

  Of course, as he’s saying this to me on the phone, I’m looking around the room, and I’m noticing, as if for the first time, the weird ways that light and shadow seem to bunch up in the corners and along the edges—authentically, I guess—I look out the window at the sky—and how many times have you looked at a sky and said, “If this were in a movie, I would criticize the special effects”?

  Should you paint

  a credible sky

  you must keep in mind

  its essential phoniness.

  –EDUARDO HURTADO

  Devon’s most recent assignment had been to work on rocket-launcher contrails, a problem that proved trickier than he’d originally thought; he stayed late many long evenings trying to get its waviness and dispersion just so. He finally nailed it, and the studio was pleased: it went into the film. But all that scrutiny came with a price. Now, when he goes outside and looks at airplane contrails, he’s suspicious. “When I was working on those smoke trail things—when I was out hiking or something, watching the planes go by, and trying to analyze how the shape changed over time … You almost, like, question reality at times—like, you’re looking at something, like smoke or something, and you think, that’s too regular, the smoke shouldn’t look so regular …”

  That’s what I keep feeling, now, when I read an email or pick up the phone. Even with my own parents—I found myself waiting, like the phonagnosic Steve Royster, for the moment they said something incontrovertibly, inimitably “them.”

  The Unsung Beauty of the Glassware Cabinet

  Curious learning not only makes unpleasant things less unpleasant, but also makes pleasant things more pleasant. I have enjoyed peaches and apricots more since I have known that they were first cultivated in China in the early days of the Han dynasty; that Chinese hostages held by the great King Kaniska introduced them into India, whence they spread to Persia, reaching the Roman Empire in the first century of our era; that the word “apricot” is derived from the same Latin source as the word “precocious,” because the apricot ripens early; and that the A at the beginning was added by mistake, owing to a false etymology. All this makes the fruit taste much sweeter.

  –BERTRAND RUSSELL

  Reflection and refraction are difficult to simulate on a computer. So is water distortion. So-called “caustics,” the way that a glass of wine refocuses its light into a red point on your table, are particularly hard to render.

  Reflection and refraction are also fairly computationally nasty because they have the habit of multiplying off of each other. You put two mirrors in front of each other, and the images multiply to infinity in no time flat. Light travels roughly 200,000 miles per second: that’s a lot of ping-pong, and way beyond the point where most rendering algorithms tap out. Usually a programmer will specify the maximum acceptable number of reflections or refractions and cap it there, after which point a kind of software deus ex machina sends the light directly back to the eye: no more bouncing.

  Getting off the phone with Devon, I go to my kitchen and fling open the glassware cabinet. I am more mesmerized by the hall of mirrors within than I have ever been before. My eyeball bulging at the side of a wineglass, I watch real life, real physics, real light perform.

  A glassware cabinet is a computational nightmare.

  So, Devon explains, is a deciduous forest. And nude bodies are more of a computational nightmare than clothed ones: all those tiny hairs, irregular curvatures, semi-translucencies of veins under slightly mottled skin.

  I love these moments when the theory, the models, the approximations, as good as they are, aren’t good enough. You simply must watch. Ah, so this is how nature does it. This is what it looks like. I think it’s important to know these things, to know what can’t be simulated, can’t be made up, can’t be imagined—and to seek it.

  Devon, in his life out of the studio, now pays a kind of religious attention to the natural world. It helps him do a better job in his animation projects, I’m sure, but one suspects the means and ends are actually the other way around.

  “It’s nice to know at least that there are quite a few things that, at least with computer graphics, and what I do, that I’m like, I mean, Wow. You know, I wrote some thing, and I have people waiting, you know, ten hours for a frame, and it doesn’t even look realistic, it doesn’t even look quite right! And I’m like, Damn—that’s, one, quite far from reality, and, two, it’s stretching, like, x number of dollars’ worth of computing at it and barely even making it. So.”

  Devon laughs.

  “It feels … It definitely feels good at the end of the day that I can open my eyes and look at something that’s, like, many orders of magnitude more complex.”

  And to be able to know where to look for it—

  And how to recognize it.

  Acknowledgments

  It was Isaac Newton who famously said (though it was actually a common expression at the time), “If I have seen a little further it is by standing on the shoulders of giants.” I want to say, more neurologically, that if I’ve been able to conduct a good signal to my axon terminal, I owe it to the people at my dendrites. (Though it goes without saying, of course, that any noise or error in the signal is my own.)

  I’m indebted to a number of conversations with friends and colleagues, which sparked or contributed many of the specific ideas in the text. I recall, in particular, such conversations with Richard Kenney, David Shields, Tom Griffiths, Sarah Greenleaf, Graff Haley, François Briand, Greg Jensen, Joe Swain, Megan Groth, Matt Richards, Emily Pudalov, Hillary Dixler, Brittany Dennison, Lee Gilman, Jessica Day, Sameer Shariff, Lindsey Baggette, Alex Walton, Eric Eagle, James Rutherford, Stefanie Simons, Ashley Meyer, Don Creedon, and Devon Penney.

  Thanks to the researchers and experts of their respective crafts who graciously volunteered their time to speak at length in person (or the closest technological equivalent): Eugene Charniak, Melissa Prober, Michael Martinez, Stuart Shieber, Dave Ackley, David Sheff, Kevin Warwick, Hava Siegelmann, Bernard Reginster, Hugh Loebner, Philip Jackson, Shalom Lappin, Alan Garnham, John Carroll, Rollo Carpenter, Mohan Embar, Simon Laven, and Erwin van Lun.

  Thanks, too, to those with whom I corresponded by email, who offered thoughts and/or pointed me toward important research: Daniel Dennett, Noam Chomsky, Simon Liversedge, Hazel Blythe, Dan Mirman, Jenny Saffran, Larry Grobel, Daniel Swingley, Lina Zhou, Roberto Caminiti, Daniel Gilbert, and Matt Mahoney.

  Thanks to the University of Washington Libraries and the Seattle Public Library; I am in your debt, quite literally.

  Thanks to Graff Haley, Matt Richards, Catherine Imbriglio, Sarah Greenleaf, Randy Christian, Betsy Christian, and, with special appreciation, Greg Jensen, all of whom read and offered feedback on an earlier draft.

  Thanks to Sven Birkerts and Bill Pierce at AGNI, for publishing an earlier version of “High Surprisal” (as “High Compression: Information, Intimacy, and the Entropy of Life”) in their pages, for their sharp editorial eyes and their support.

  Thanks to my agent, Janet Silver at Zachary Shuster Harmsworth, for believing in the project from day one, and for her support, wisdom, and enthusiasm throughout.

  Thanks to my editors, Bill Thomas and Melissa Danaczko, and the rest of the Doubleday team, for their expert eyes, and for all of the faith and hard work of
bringing the book into the world.

  Thanks to invaluable fellowships at the Bread Loaf Writers’ Conference in Ripton, Vermont; at Yaddo in Saratoga Springs, New York; and at the MacDowell Colony in Peterborough, New Hampshire. A reverence for good work fills them with a kind of airborne sacredness like very few places I know.

  Thanks to the baristas of Capitol Hill and Wallingford for the liquid jumper cables of many Seattle mornings.

  Thanks to a hamster-sitting residency at the Osborn/Coleman household, where good work was done.

  Thanks to Michael Langan for a very fine portrait.

  Thanks to Philip Jackson, for allowing me to be a part of the 2009 Loebner Prize competition, and to my fellow confederates, Dave Marks, Doug Peters, and Olga Martirosian, with whom I was proud to represent humanity.

  Thanks to my parents, Randy Christian and Betsy Christian, for the unconditional everything along the way.

  Thanks to the inestimable Sarah Greenleaf, whose clarity of mind cut many a Gordian knot, and whose courage and compassion have shaped both the text and its author.

  Thanks to everyone who has taught me, by words or by example, what it means to be human.

  Notes

  Epigraphs

  1 David Foster Wallace, in interview with David Lipsky, in Although of Course You End Up Becoming Yourself (New York: Broadway Books, 2010).

  2 Richard Wilbur, “The Beautiful Changes,” The Beautiful Changes and Other Poems (New York: Reynal & Hitchcock, 1947).

  3 Robert Pirsig, Zen and the Art of Motorcycle Maintenance (New York: Morrow, 1974).

  4 Barack Obama, “Remarks by the President on the ‘Education to Innovate’ Campaign,” press release, The White House, Office of the Press Secretary, November 23, 2009.

  0. Prologue