Gene Roddenberry bought his first computer in 1983. It was a Kaypro 10. It was a clunky metal box, the size of a microwave oven. The front panel came off to become a keyboard, revealing a 9-inch monochromatic screen that glowed with unholy green alphanumeric characters. Inside, it had a 4MHz Z80 chip, 64K of RAM, and a 10-megabyte hard drive. Its operating system was CP/M.
Knowing that I had some skill with computers, Gene called me and asked for help. So I gathered up several boxes of 5 1/4-inch floppies (they really were floppy then) and drove to Beverly Hills. (To this day, I still think of Beverly Hills as a very strange foreign country.)
At that time, Gene’s career was in a slump. Paramount Pictures had not been happy with how the budget of “Star Trek: The Motion Picture” had ballooned out of control, and they had ungently pushed Gene aside and handed the franchise to Harve Bennett. Gene wasn’t happy about that, and he worked mostly at home for a few years. Upgrading from a typewriter to a computer was a way for him to keep up with the future.
I taught Gene a lot of the necessary basics: how to boot a program, how to save, how to copy—things we take for granted nowadays but were arcane mysteries to most people in the earliest days of microcomputing. I also gave him a couple of games, because nothing reduces your fear of technology like turning it into a toy.
One of the games I gave him was Eliza. Eliza was a chatterbot program, originally created in 1966, but easily ported to microcomputers. Wikipedia describes Eliza as “a computer program designed to simulate a therapist or psychoanalyst” but more than that, it was an opportunity to experience human-computer interaction.
From a programming perspective, Eliza was a simple exercise, simple enough to be coded in a few hundred lines of BASIC, but very clever in the way its algorithms mimicked actual conversation. For instance, if you typed, “I don’t like [predicate clause],” the program would recognize “I don’t like” in the string and replace it to respond, “Why don’t you like [predicate clause]?”
If the Eliza program couldn’t find a question to ask, it would say, “Tell me more,” or it would ask an open-ended question. Eliza was a psychological mirror. Conversations with her could be funny, wistful, sad or even disturbing. Some people found it easier to talk to Eliza than to another human being. (Myself, I found conversations with Eliza to be uninteresting, because Eliza had nothing of its own to contribute. It felt like a passive-aggressive alien intelligence—and who needs a virtual anal probe anyway?)
But this was Gene Roddenberry’s first exposure to Eliza, and he was delighted by the (simulation of) conversation. At one point in the afternoon, Gene’s wife, Majel, came home from wherever she had been and looked in on us. Gene told her, “David and I are computing.” To this day, that still strikes me as a very odd sentence. Computing what?
“Here,” said Gene. “Look at this!” He sat Majel down at the keyboard and told her to type “Hello.” She did and the Eliza program responded with, “Hello, tell me something about yourself.” And off they went.
Majel was astonished—even a little freaked out. At one point, she backed away from the computer and shrieked, “Gene, who’s in there?”
I had to explain that it was a simulation, not a real intelligence, but I’m not sure Majel ever believed me. In those early days of home computers, Eliza was startling.
Ironically, Majel had provided the voice of the computer on the starship Enterprise. Scriptwriters often had Kirk turn to the computer for necessary exposition. He’d ask a question, and the computer would respond (somewhat mechanically): “Working…” and then a short moment later, “The airspeed velocity of an unladen swallow is…” or whatever other information was needed to push the story forward.
Yesterday, I had a story meeting. I’d forgotten to check the address, so I tapped my Android phone and said, “Navigate to Octopus restaurant, Encino.” The phone showed me the addresses of three restaurants named Octopus. I tapped the Encino address and the phone immediately began giving me instructions on the route: “Turn right on Balboa Blvd. in 500 feet.” And so on.
Some people think this is spooky. Some people interact with their GPS tools as if there’s a troublesome imp inside. I don’t. I understand that it’s merely a collection of useful algorithms, sorting and selecting and reporting. This is Eliza, all grown up. Add a connection to Google’s mainframe, some sophisticated speech recognition and a massive database, and you have the first useful iteration of the Enterprise’s own computer. (And in a delicious bit of life imitating art, Google’s speech-recognition system is code-named Majel.)
Someday, perhaps, we may be building intelligence engines as sophisticated as Arthur C. Clarke’s HAL 9000. Neither Apple’s Siri nor Google’s Majel is anywhere near that point yet, because neither are intelligent in the human sense. Both are still rule-based algorithms, reporting the most appropriate answers that can be gleaned from their data stores.
True machine intelligence is still more theory than practice. It touches on questions that we cannot yet answer. What is consciousness? Why do we experience ourselves as selves? And if we cannot answer those questions, how can we create software that effectively behaves with real intelligence?
What do you think?
David Gerrold is the author of over 50 books, several hundred articles and columns, and over a dozen television episodes, including the famous “Star Trek” episode, “The Trouble with Tribbles.” He is also an authority on computer software and programming, and takes a broad view of the evolution of advanced technologies. Readers may remember Gerrold from the Computer Language Magazine forum on CompuServe, where he was a frequent and prolific contributor in the 1990s.