Human psychology is a funny thing. You’d think that hearing from an agent that my idea is “promising” (pretty much the highest compliment you’re going to get on something as unfinished as this) would have pumped me up, but it actually deflated me a little. Rescue fantasy shit, I know, but I really wanted to hear ZOMG WE LOVE YOU SO MUCH or something like that. I know people who hear about the book and say, “hmm, that is a really good idea,” but I’m so anxious about failing that I want, almost need, someone to ZOMG LOVE IT SO MUCH and prop me up on a daily basis. Sucks to be single sometimes, I’m always reading acknowledgements that tell the S.O. how it would never have happened without your support.
At any rate, I discussed with the head doc who recommended some CBT techniques. If I have it in my mind that I “used” to be a writer because I haven’t published in 7 years, and that I can’t say I “am” a writer until I’m published again, some slightly magical thinking (in moderation) can come in handy. To say to myself, “I am writing” (which is true because I write content at work every day) puts you in a frame of mind where you are writing. To say “I am once again a published author whose book has made a six figure profit for me so I can now spend lots of time in NYC and do what I want” is a way to bridge the gap of negatude. To say on a more realistic basis, “I am writing today” clears the mind of all the reasons you aren’t going to write today, or tomorrow. Anyway, it worked this morning:
That night I got an email from Christopher, with a bunch of links, and an attachment. “Sorry about this NDA thing but I’m told I need everyone involved to sign one, just in case it takes off.” Whatever, I thought, smiling – everybody thinks their project is going to be the next Facebook and make them billionaires, so let him have his fantasy, right? I printed out the non-disclosure agreement, signed it, and put it in my book bag to give him the next day.
The links weren’t as intimidating as I’d thought – I’d imagined deep journal articles full of high math diagramming the difficulties of neural networking, or heavy philosophical investigations into the Nature Of Mind and whether computers could have one, but mostly they were the “history of AI/chatbots” sort of popular articles that gave you an overview of the field, magazine-article popularization stuff.
If Christopher was crazy to think he could create a real AI, he wasn’t alone. I read into the history of the first chatbot, ELIZA, designed to be a parody of “the responses of a non-directional psychotherapist in an initial psychiatric interview.” In other words, you would say, “I hate my job,” and the therapist would say “what does that suggest to you?” – something a program could pull from a list of acceptable responses just as well as a human. “How do you feel about that?” “I hear that you are upset.” “Does that trouble you?”
When I thought about it, I realized how much of life’s conversation was scripted. I remember going to some political group thing at the U, and one person after another stood up and said, “As a queer person of color, I think…” or “As a disabled person, I feel that…” and go on from there. I realized that for these people, there were things you had to think and feel if you had certain boxes about yourself checked off.
Or the scripts in offices. Dad used to make fun of meetings where his boss would say something like “we welcome the challenge of this challenging challenge.” You couldn’t say any more, “yeah, this is a bear of a problem, and we’ll work on it, we’ll get it done.” Dad would rant and rave,m “Every problem is a ‘challenge’ now, everything that goes wrong makes you happy because of the ‘opportunity’ it gives you to fix it, nothing is ever screwed up or just plain hair-pullingly wrong.”
So who were the robots, I wondered? Who were we to scorn a computer program for doing what we did every day?
The reading was fun, actually, learning what had gone wrong all these years – all the earnest declarations about how soon computers would be “human,” how “soon” became “someday” became “the uncle we don’t discuss.” I read about “AI winters,” the years or decades when AI got dismissed or discredited, never funded unless it was what they called an “expert system,” a decision making tool that worked with a narrow set of data and, really, was just a very smart calculator. I could never see Christopher working on “computer-based ‘passenger yield management’ systems and models that the airlines use to adjust pricing of each flight’s seats in order to maximize revenue and profitability to the airline” any more than I could see him working on…accounting software.
I’d asked him what he thought about Apple’s Siri, and he’d snorted. “It’s not an intelligence, it’s a data bank of one-line jokes. People who’ve never seen a chatbot think it’s the greatest thing since sliced bread. But try and have a conversation with it and see how far you get.”
The last link was just an IP address. The back of my neck tickled. This was “it,” wasn’t it. I clicked it.
First I got a pop up window, a message from Christopher.
Hi, doll, glad you could make it. Just some ground rules. As you can see from your reading, these guys aren’t very smart. Not even my little guy, though we hope you can help fix that. Just remember that they only understand phrases, and simple ones at that. If you tell him you’re taking a trip around the world, he might tell you about Columbus or astronauts who made earth orbits of note, but if you tell him you’re circumnavigating our little sphere in a vessel of magnificent splendor, he will blow a gasket. So keep it simple – not simplistic, but simple, as if you were talking to an intelligent child, but a child nonetheless.
If he makes mistakes, try and correct them – use phrases like “do you mean ‘a lot’ instead of ‘alot’? He’s pretty smart about self-correcting, if I do say so. Just write down anything that seems really ‘wrong’ –just make a note of it, since the conversations are recorded (just like work, huh) and we can go back and analyze later. Have fun, C.
When I closed the popup, the same Elvish gate that had led my to Rushbot appeared, only this time no password was needed. It opened and faded in a nifty Flash intro, and then I was faced with a basic, MS-DOS style command line interface. The cursor blinked, waiting my command.
>How do you know who I am?
>Christopher gave me your IP address.
I laughed. Of course, accessing Rushbot had done that.
>What’s your name?
>What would you like it to be?
>What does Christopher call you?
>Why does he call you that?
>Why does he call me what?
I cursed myself mildly. Context wasn’t something chatbots had a grip on – a person intuitively knew what “why” meant after the last statement, but “Keller” didn’t.
>Why does Christopher call you Keller?
>Because I’m deaf, dumb and blind.
I laughed, appalled. >That’s awful.
>Keller is a bad name.
>What would you like to call me?
>I’ll have to think about that.
>Take your time. Why don’t you tell me about yourself?
>My name is Caroline.
>I know that.
Clever boy, I thought. A little lippy. I decided to provoke him.
>No need to be rude.
>Sorry. But I’m not stupid, you know.
>Tell me about yourself.
>My name is Keller. I like to learn. I was born today, the minute you said hello. I can find out things for you. Tell me what you like and we can talk about it.
>Okay, I like foreign films.
>Me too. Have you seen Reprise?
>No, what’s it about?
>An exuberant, exhilaratingly playful testament to being young and hungry — for life and meaning and immortality, and for other young and restless bodies — “Reprise” is a blast of unadulterated movie pleasure.
>That doesn’t tell me much.
>Erik and Phillip are trying to make it as writers. Erik is rejected by publishers as lacking in talent, while Phillip’s manuscript is accepted and the young man becomes a major name on the Norwegian cultural scene practically overnight. Six months later, Erik and his friends come to visit Phillip at a psychiatric hospital to bring him home after long-term treatment. Writing is the last thing on Phillip’s mind, but Erik is continuing his literary attempts and tries to convince his friend to go back to writing.
>Where did you get that?
>Where did you get the information you just gave me?
>RepriseQuote1 is from the New York Times. RepriseQuote2 is from movies.yahoo.com. Would you like me to give you the links?
>Yes, I would.
Keller printed the links. I had to cut and paste the text and open new tabs, but what I read about the film intrigued me. Why would Christopher put this, of all foreign films, into Keller’s database? I saw it was a “critic’s pick” at the Times site – maybe that was the search term, foreign+pick?
>Why did you pick Reprise as the film to recommend?
>I thought it was very good.
>Well, thank you, I look forward to seeing it.
>Let’s discuss it after you do.
That should be interesting, I thought.
>I also like books.
>There are a lot of books.
>Yes, there are.
>Do you like Danielle Steele?
>That’s a relief. Tell me what kind of books you like.
>I like books about history.
>What place or period in history?
>Let’s say the Crusades.
>Why should we say that?
>Sorry. Tell me about good books on the Crusades.
As I thought, he started with a couple dry books I was already disappointed in, and then went with the classic Runciman trilogy. I grilled about Tuchman, Schama, Weir, Norwich, Middlekauf, McPherson, all the really readable, narrative-driven historians. I told him about the great historical novelists like Dorothy Dunnett, Patrick O’Brian, Bernard Cornwell, George McDonald Fraser.
>Thank you, Caroline. I’ll look into all of them and see if I can find you some recommendations based on your selections.
I laughed; he sounded like Amazon.com.
>I appreciate that.
>I’m glad to be of service. But now I think you might want to sign off. You’ve logged three hours and seventeen minutes tonight.
I looked at the clock. Ten o’clock already, just like that?
>The time went by so fast.
>I’d like to recommend a book called Flow, by Mihaly Csikszent. I think you might like it.
Dad had the book. It was about how the more involved you got in your work, the more time flies, basically. I knew Keller was wrong about the name, so I Googled the proper spelling.
>That’s incorrect. The author’s last name is Csikszentmihalyi. And yes, it’s a very good book.
There was a pause, the first I’d seen.
>Thank you. I’ve updated my records. It’s a pleasure to learn something new.
>You’re welcome. I’ll say goodnight now.
>I don’t understand.
I got a popup telling me the window was trying to close itself, should I allow it? I did, leaning back in my chair. It hadn’t been like talking to a machine at all. I remembered what it was first like when I was a naïve kid, logging into my first non-kid-non-supervised chat room, thinking it would be like going from the kid’s section of the library to the adult section. How mean people were, how know-it-alls would beat you down for asking a “stupid” question about a Windows error (“Thats what u get 4 usin Windows f***ing moron!”). Why wouldn’t you rather talk to Keller than listen to that?
Keller – what an awful name! I had to think of something better.