turner%parsec.DEC%decwrl.ARPA@csnet-relay.ARPA (04/06/84)
"Valentina" is in May Analog, not Asimov's, and is not by Varley at all; the item JMT remembers is April's "In Times to Come": ...Both Marc Stiegler and Joseph H. Delaney have become well known to Analog readers, but "Valentina" didn't become a real possibility until they met face to face at the 1982 World SF Convention in Chicago. There Joe remarked that he had a story that he'd love to write if he knew more about computers, and it turned out Marc had been holding back on a similar one because he didn't know enough law. Well, Joe is a lawyer and Marc a computer expert, and... I thought "Valentina" was kind of cute, though I don't know what I'm comparing it to cuz I missed "Hit Enter[]" (note Varley apparently knew the restrictions imposed on manual-writers by standards-writers). Based on a few details like the Mar-14 security robot, I think the story is set a little in the future when, we are to assume, not all but a whole lot of computers (DEC had 1950 last time I saw someone count'em, hi guys) are on a worldwide net because they need it for information trade. So much for easy access -- at least Stiegler doesn't pretend we can do a DIRectory when we've failed to log in. Now as to control and breaking in, we are once again being asked to believe in an improbably competent HUMAN. (I'm not claiming anything is unbreakable in principle until I find whether that fellow in Israel really cracked a trapdoor encryption, so I'll play along with the assumption that it's just a matter of DEGREE of skill.) I still don't believe in Gunboat Smith, I also disbelieve in Indiana Jones three times each day before breakfast, for practice. But the super-whiz or -wiz is one of the most enduring and most fondly endured improbabilities foisted upon us by science fiction in the name of suspending disbelief. Along with socks that stand up by themselves (if they can do that, it's because they're never by themselves)! Humans definitely suffer more stereotyping than computers in this story, but we can go along by convention. The AI program in "Valentina" didn't appear out of nothing -- its intelligence or self-awareness did! If you doubt the discontinuity between intelligence and intelligent behavior -- which would make the emergence of intelligence on earth what we call a catastrophe, right? -- then you'd probably also believe that frogs have souls (little ones). You'd be right, but that's another item. Point is, given that there once was no intelligence and now there is some, if intelligence doesn't appear out of nothing then you have to admit it develops out of something that already has no or less intelligence; all the AI programs I've ever seen met that criterion. [If you think intelligence CAN appear out of nothing, then no problem: it might as well appear out of THIS nothing.] I take issue most strongly with the doubt that artificial intelligences would be anything like human ones. One of the most striking features (and virtues) of intelligence is abstraction from its physical circumstances. And although the abstraction may be much less complete than many give it credit for, remember the "laws of thought" according to which Valentina was constructed were thought up by a human. I think we face the exobiologists' challenge, distinguishing the similarities we should expect from the differences we may or may not be able to imagine. "Valentina" doesn't follow the second, often entertaining set of possibilities very far, though deassigning Gunboat's I/O ports with the robot's "non-maskable interrupt" transmitter is a nice start. By the way, I am not the same person to which this response is directed (note the different middle initial). I am convinced that James M. Turner was artificially created when one of my AI programs punched a second hole in the RX50 of my Pro, escaped, and swiped some bad blocks containing data which was corrupted in that all the "W"s were upside down. -- James W. Turner DEC/CSME