Terminal Compromise

Pierre found the analogy worth a prolonged fit of giggles through which he managed to ask, "but what if you told the computer what it meant and it learned from there. On its own. Can't a com- puter learn?"

Max was seriously stoned. "Sure I guess so. Sure. In theory it could learn to do your job or mine. I remember a story I read by John Garth. It was called Giles Goat Boy. Yeah, Giles Goat Boy, what a title. Essentially it's about this Goat, musta been a real smart goat cause he talked and thunk and acted like a kid." They both roared at the double entendre of kid. That was worth another joint.

"At any rate," Max tried to control his spasmodic chuckles. "At any rate, there were these two computers who competed for control of the world and this kid, I mean," laughing too hard to breath, "I mean this goat named Giles went on search of these computers to tell them they weren't doing a very good job."

"So, what has that got to do with an Apple learning," Pierre said wiping the tears from his eyes.

"Not a damn thing!" They entered another spasm of laughter. "No really. Most people either think, or like to think that a com- puter can think. But they can't, at least not like you and me. " Max had calmed down.

"So?" Pierre thought there might still be a point to this conver- sation.

"So, in theory, yeah, but probably not for a while. 10 years or so."

"In theory, what?" Pierre asked. He was lost.

"In theory a machine could think."

"Oh." Pierre was disappointed.

"But, you might be able to emulate thinking. H'mmmm." Max re- treated into mental oblivion as Abbey Road played in the back- ground. Anything from Apple records was required listening by Max.

"Emulate. Emulate? What's that? Hey, Max. What's emulate? Hey Max, c'mon back to Earth. Emulate what?"

Max jolted back to reality. "Oh, copy. You know, act like. Emulate. Don't they teach you emulation during sex education in France?" They both thought that that was the funniest thing ever said, in any language for all of written and pre-history.

The substance of the evening's conversation went downhill from there.

A few days later Max came by Pierre's loft. "I been thinking."

"Scary thought. About what?" Pierre didn't look up from his Apple.

"About emulating thought. You know what we were talking about the other night."

"I can't remember this morning much less getting shit faced with you the other night."

"You were going on and on about machines thinking. Remember?"

"Yes," Pierre lied.

"Well, I've been thinking about it." Max had a remarkable ability to recover from an evening of illicit recreation. He could actually grasp the germ of a stoned idea and let a straight mind deal with it the following day. "And, I maybe got a way to do what you want."

"What do I want?" Pierre tried to remember.

"You want to be able to label all of your music so that to all appearances each piece of music knows about every other piece of music. Right?"

"Kinda, yeah, but you said that was impossible . . ." Pierre trailed off.

"In the true sense, yes. Remember emulation though? Naw, you were too stoned. Here's the basic idea." Max ran over to the fridge, grabbed a beer and leapt into a bean bag chair. "We assign a value to every piece of music. For example, in music we might assign a value to each note. Like, what note it is, the length of the note, the attack and decay are the raw data. That's just a number. But the groupings of the notes are what's important. The groupings. Get it?"

Pierre was intrigued. He nodded. Maybe Max did understand after all. Pierre leaned forward with anticipation and listened intent- ly, unlike in one ear out the other treatment he normally gave Max's sermons.

"So what we do is program the Apple to recognize patterns of notes; groupings, in any size. We do it in pictures instead of words. Maybe a bar, maybe a scale, maybe even an entire symphony orchestra. All 80 pieces at once!" Max's enthusiasm was conta- gious. "As the data is put in the computer, you decide what you want to call each grouping. You name it anything you want. Then we could have the computer look for similar groupings and label them. They could all be put on a curve, some graphic of some kind, and then show how they differ and by how much. Over time, the computer could learn to recognize rock'n'roll from Opera from radio jingles to Elevator Music. It's all in the patterns. Isn't that what you want?" Max beamed while speaking excitedly. He knew he had something here.

Max and Pierre worked together and decided to switch from the Apple II computer to the new IBM PC for technical reasons beyond Pierre's understanding. As they labored, Max realized that if he got his "engine" to run, then it would be useful for hundreds of other people who needed to relate data to each other but who didn't know much about computers.

In late 1982 Max's engine came to life on its own. Pierre was programming in pictures and in pure English. He was getting back some incredible results. He was finding that many of the popu- lar rock guitarists were playing lead riffs that had a genealogy which sprang from Indian polyphonic sitar strains.

He found curious relationships between American Indian rhythms and Baltic sea farer's music. All the while, as Pierre searched the reaches of the musical unknown, Max convinced himself that everyone else in the world would want his graphical engine, too.

Through a series of contacts within his Big Eight company, Max was put in touch with Hambrecht Quist, the famed Venture Capital firm that assisted such high tech startups as Apple, Lotus and other shining stars in the early days of the computer industry. Max was looking for an investor to finance the marketing of his engine that would change the world. His didactic and circumlocu- tous preaching didn't get him far. While everyone was polite at his presentations, afterwards they had little idea of what he was talking about.

"The Smart Engine permits anyone to cross-relate individual or matrices of data with an underlying attribute structure that is defined by the user. It's like creating a third dimension. Data is conventionally viewed in a two dimensional viewing field, yet is really a one dimension stream. In either source dimensional view, the addition of a three dimensional attribute structure yields interrelationships that are not inherently obvious. Thus we use graphical representations to simplify the entire process."

After several weeks of pounding the high risk financial community of the San Francisco Bay area, Max was despondent. Damn it, he thought. Why don't they understand. I outline the entire theory and they don't get it. Jeez, it's so easy to use. So easy to use. Then the light bulb lit in his mind. Call Pierre. I need Pierre. Call Pierre in New York.

"Pierre, it's Max." Max sounded quite excited.

"How's the Coast."

"Fine, Fine. You'll find out tomorrow. You're booked on American #435 tomorrow."

"Max, I can't go to California. I have so much work to do."

"Bullshit. You owe me. Or have I forgotten to bill you for the engine?" He was calling in a favor.

"Hey, it was my idea. You didn't even understand what I was talking about until . . ."

"That's the whole point, Pierre. I can't explain the engine to these Harvard MBA asswipes. It was your idea and you got me to understand. I just need you to get some of these investors to understand and then we can have a company and make some money selling engines." Max's persistence was annoying, but Pierre knew that he had to give in. He owed it to Max.