2001: HAL

[Steve Jordan]

Don't forget: The movie was designed to be a movie. Presenting HAL as a bunch of solid circuit modules that could be unplugged and left floating about made for good cinema, however accurate it was realistically. (Same goes for the message from Dr. Floyd that starts up, in the computer's memory room, after Bowman pulls the modules.)

As I remember the book, after Bowman shut down HAL's higher functions, he had his hands full keeping the ship running properly, taking over at least part of HAL's job. This may not seem to make sense on the surface, but it seems to fit established spacecraft procedures of the time, using automation where possible, but leaving it to the astronauts to be back-up for downed automation systems.

The open question is, exactly how much of the ship's systems was HAL running autonomously, and how much was part of the higher functions? Clarke isn't specific on that, as well he should not have been... that would have been simply too much extraneous detail for the story (though I suspect it's something that modern authors would be advised to wax eloquently on for 10-20 pages, to help pad out a story...).

[kythe]

I find Hal to be one of the most intriguing aspects of the story. I do think he could feel real emotions, in that he was an entirely sentient being. He exceeded his intended programming.

In an sense, his "evolution" parallels that of the human race in the stories. Hal starts out as a computer and, like the ape-men, he is altered by a higher species until he becomes like them in terms of true intelligence and sentience. Unlike humans, though, his "creators" were not prepared for the level of his advancement and shut him down.

Throughout the book, there are several discussions about leaving the flesh behind. First, this happens through uploading consciousness into a machine, then eventually by shedding the machine and becoming pure consciousness. I think Hal is a conscious being, but he bypassed he "flesh" stage in that he was created as an AI. He also has potential for evolution, but has no one to lead him.

I see Hal as the most tragic character in the story. He is the least understood and very much alone and without guidance. Yet, he is more "human" than anyone else in the story.

[j. d. worthington]

Very interesting take. I like that, Kythe. Um, sure you're not related to Dr. Chandra?

[kythe]

I don't know. Who's Dr. Chandra?

[j. d. worthington]

Originally Posted by kythe

I don't know. Who's Dr. Chandra?

Here you go:

Dr. Chandra - Wikipedia, the free encyclopedia

[kythe]

Ah, I see. This means I need to read "2010" now. After I see the movie version of Space Odyssey: 2001, that is.

[j. d. worthington]

Originally Posted by kythe

Ah, I see. This means I need to read "2010" now. After I see the movie version of Space Odyssey: 2001, that is.

2010 is a good book (the movie has its moments, too, for that matter, though it's nowhere near as powerful a visual experience as Kubrick's film -- which should be seen on as big a screen as possible, by the way); but 2061 really isn't worth bothering with, in my opinion....

[mosaix]

Originally Posted by AaronAgassi

Perhaps HAL's components are hot swappable, enabling maintenance during mission critical operation, uninterrupted.

Quite right. Fault tolerant systems demand hot swap components and there's no reason why this can't extend to such things as logic and memory boards.

I have to disagree with HAL feeling emotions however. Computers have very simple instruction sets - add, subtract, multiply, divide, copy, read, write, compare, goto (or branch), some kind of interrupt and halt. There are variants of these commands but essentialy that's it. These commands are used, at high speed, to give the impression of intelligence.

The computer is running, per processor, one, just one of these commands at any moment in time. Out of the above list just which command could be feeling an emotion?

With clever programming it can appear that a computer can be intelligent, have emotions and almost be human - but it can't it's just clever programming.

[j. d. worthington]

I'm not quite sure I agree with you on this, mosaix. Mind you, it's been a very long time, but my impressions and memories were more along the following lines (bear with me on this):

We weren't really given much of the theory or working behind HAL, for one thing, and I'd always got the impression these were the result of a divergence in early cybernetic theory, resulting in something that was what one might call a "threshold"... somewhere between machines as we understand them and a genuine artificial intelligence; in part because of extremely complicated interconnecting (and proliferating in interconnections) linkages in a cybernetic "neural web" (for lack of a better term). After all HAL was capable of not only learning from experience, but also of making a certain level of judgment calls, which would imply not only the weighing of facts, but the assigning of values, especially where its human counterparts were concerned.

Also, the confusion HAL suffered from the conflict of its basic programming versus the overlain programming which caused it to lie and hide facts from its human "colleagues" ... which it was nonetheless to treat with perfect candor in every other way ... caused something which was handled very like a genuine nervous or mental breakdown. It simply couldn't make these contradictory things match. It always struck me much like the intelligence and knowledge of a supergenius combined with the emotional experience and sophistication of a very young and idealistic child....

[mosaix]

I was getting a bit carried away there JD, I forgot we were talking about fiction.

I was having a little rant at the quite common mistake that people have of assuming that artificial intelligence is anything but artificial, but genuine.

[j. d. worthington]

Originally Posted by mosaix

I was getting a bit carried away there JD, I forgot we were talking about fiction.

I was having a little rant at the quite common mistake that people have of assuming that artificial intelligence is anything but artificial, but genuine.

Ah. Okay. My mistake....

Errr, that was my mistake, yes?

[Steve Jordan]

Most of the problems inherent in establishing HAL's state of being stem from a lack of understanding of the definitions themselves. Is there a real difference between "artificial intelligence" and biological intelligence? Is Intelligence less so, because it is "artificial"? Are emotions anything more than learned or instinctive mental tools used as survival traits? Does intelligence automatically qualify as sentience, or does self-awareness define sentience? And does intelligence, emotion and sentience in combination determine a "soul," or is there something else needed for that?

If intelligence is intelligence, whether naturally evolved or constructed by a second party... If emotions are applied by the intelligence to improve its interaction with others, and its survival... if self-awareness defines sentience... then by those measures, HAL is a much a sentient being as the astronauts, and his end just as tragic, because it is preordained... he is literally preprogrammed to fail. If anything, it is harder to tell, because HAL does not respond the way we do, does not demonstrate emotion the way we do, so we have no reference with which to qualify his sentience, his emotions, his soul.

However, if "artificial intelligence," by definition, doesn't count as real intelligence... if emotions and sentience can only be part of a mind with a soul... then HAL was a toaster that was unplugged.

Kubrick understood this. His film portrayal of Poole and Bowman as nearly emotionless, mirrored HAL's interactions with the astronauts. Frank's anger at HAL trying to kill him was mirrored by HAL's desperation to stop Frank from shutting him down. The deaths of the sleeping crewmen were treated as abstracts, killing men that already looked dead. Dave was literally swatted away by HAL, and when Frank could not return with the body, he was unceremoniously discarded.

There was no discernable difference between the intelligence, sentience, emotion, or soul of the astronauts, and of HAL. Kubrick did this deliberately, to blur the line between human and computer "sentience" and force the viewer to consider those definitions against machines, and against themselves.

[chrispenycate]

Heuristically programmed, algorithmically organised.
Does the lack of hormones make genuine intelligence impossible? The lack of "feelings" ie. the increase in the importance of absolute logic, render thought so divergent from the hominid norm that it is no longer recognised as thought at all, but some watered-down simulation?
Thus, by analogy, those of us who live less in our endocrine system and more in the nerve network are not really thinking; we're calculating, and camouflaging ourself amoungst our more emotive bretheren by pretending to the same innaccuracies.
Biological bases for thought processes are no more complex than gates and inverters, shift registers and RAM. and are organised considerably less logically than computer elements; in particular, programming is a lot more haphasard. Where an ant or a goldfish wins out is in sheer number of synapses. "Choice" routines (when nothing seems to give an optimum result, do something at random, rather than doing nothing and waiting for orders, or the situation to clarify)have been available for decades in chess-playing programs; ultimately suitanle for military connand robots (and to be left out of civil servant models, who should await further instructions while being torn apart by the raging mob)
Is "throwing the dice" (a random number generator) that much less sapient than "he reminded me of my cousin's boy, so I chose to save him, rather than another.
Or, if you want learned illogic, how about Rudi Rucker's boppers (software, wetware)? Given a sufficiently complex system, the results will be unpredictable. Which is what you were aiming for, no?