2001: HAL

[dwndrgn]

Ok. Passage reread. I still have the feeling that it is logically not correct but it isn't this huge gaping hole I thought it was originally. I was probably just feeling a bit pugnacious and all it wanted was an excuse to put the thing down

*************SPOILERS*****************

So here's the thing. Lightyears away from Earth, Dave has just discovered that the supercomputer running his spaceship has gone haywire - it has actually killed the four other human members of the crew. He goes to attempt to fix this without disabling the computer altogether because it is what keeps the ships functions going. In order to do this he physically goes to the main unit and pulls out the 'logic circuits'. To me that seems completely impossible. You'd disable the whole thing if you didn't reprogram it to accept the missing components or bypass them or whatever, at least that is my feeling on it. I'm not a computer programmer or a hardwire expert so I could be way off base but it just doesn't seem right to me.

[I, Brian]

Originally Posted by dwndrgn

*************SPOILERS*****************

So here's the thing. Lightyears away from Earth, Dave has just discovered that the supercomputer running his spaceship has gone haywire - it has actually killed the four other human members of the crew. He goes to attempt to fix this without disabling the computer altogether because it is what keeps the ships functions going. In order to do this he physically goes to the main unit and pulls out the 'logic circuits'. To me that seems completely impossible. You'd disable the whole thing if you didn't reprogram it to accept the missing components or bypass them or whatever, at least that is my feeling on it. I'm not a computer programmer or a hardwire expert so I could be way off base but it just doesn't seem right to me.

I think the problem is basically one of being dated - electronic devices then were effectively physical objects all lumped together - you "reprogrammed" machines by pulling bits out. Of course, these days, you'd expect some degree of software rewriting - but in the mid-1960's I don't think that computer software as we have it now was particularly envisaged.

My impression, anyway.

Besides...if computers could only be as simple as that...

Perhaps there is also the presumption that the AI components of HAL were merely riding on top of the existing system and could be removed without affecting it.

And new thread for this topic.

[polymorphikos]

My bet is with the latter suggestion.

[Foxbat]

Perhaps there is also the presumption that the AI components of HAL were merely riding on top of the existing system and could be removed without affecting it.

I would agree. I remember when Televisions were valve driven and nowadays it's solid state. But if you take a closer look inside you will find a series interconnected of circuit boards. Rather than change a damaged component, it is simply the case of replacing the board containing the damaged component (a simple plug-in task). I see HAL's logic circuits as being analagous to this.

[dwndrgn]

Dang. Shot down by computer geeks (this is not meant to be offensive to any computer geeks out there - I happen to have a great deal of respect for people who can truly understand and work with the blasted things - just jealous I expect )

I can see where early computers would have been created differently - more hardware driven than software driven. However, seeing Foxbat's info on the tv setup makes me even more convinced that you couldn't have separated the AI components (if HAL were a real computer) so simply. A tv set doesn't have to complete logic circuits and reasoning. It's all electricity here, switch there, yada, yada, yada. For a complex computer, AI would have to be integrated with EVERY component of the computer or it wouldn't have worked properly. The basis of AI (IMO) is that the regular computations that a normal computer would make are enhanced by human-like reasoning. This would mean that the AI components would have to be intertwined with all the basic logic circuits and wouldn't be just lying on top.

And one last note, this is all just speculation considering that something like HAL doesn't currently exist and may never do so

[I, Brian]

Actually, I wouldn't be surprised if AI isn't limited to a specific central processing core, which can be added to an existing system - much the way in that computer users can add a graphics card, soundcard, or various other fillers for their PCI slots.

Though, of course, I'm sure there's an alternative possibility of exactly as suggested, with the needs of the aI system written into the actual general operating system.

[Foxbat]

Perhaps the TV was not a good analogy.

Consider this: I am thinking about my reply to your post. I write my reply to your post. Throughout all this action, my heart continues to beat - and yet I expend none of my thought process to this task. It is an involuntary action. Yes, you could argue, that your very intelligence is connected to the fact of my heartbeating but, if I were to suffer some kind of trauma which affected my self-awareness, my heart would still keep on beating.

This is how I see the levels within HAL's AI.

[polymorphikos]

A person can be lobotomized and become a vegetable, devoid of higher brain functions, and yet still survive physicaly.

[willb]

Yes, so what Dave does is a lobotomy on Hal. With current technology you can 'hot swap' hard disks and other peripherals; while these aren't exactly essential to the running of the system, it shows that in principle parts of a computer can be added and removed with no effect on the core running parts.

But, there must have been a good reason for Hal to be included on the ship (it's been a while since I read it/saw the film, so I can't remember exactly what his function was) and if Dave killed Hal then surely he would have a great deal of difficulty maintaining all the functions Hal was there for.

[dwndrgn]

The main reason that HAL (as an AI computer instead of just a 'regular' computer) was on the ship was to be the last resort crewmember - should anything happen to the biological crew, mr. mechanical crew would take over so that no matter what, images and information regarding the portion of the solar system the ship was travelling through and to would get back to earth. The ship itself was never expected to go back to Earth so the images and info had to continue to be collected and sent back.

[steve12553]

In something as critical as a computer integrated into a spaceship maybe the idea of components is the best. I never wanted a TV/DVD player or a cmarea/phone because if one goes down you have to send in both for repairs. Keeping the AI separate or layered above the life support is a wonderful idea and the kind that NASA would had used if there still was really a space program.

[AaronAgassi]

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

[manephelien]

Have you seen the movie? It's shown pretty well there, David Bowman just pulls out loads of circuit cards, looking a lot like today's graphics and soundcards, until HAL's AI is disabled, while retaining automatic functioning.

[j. d. worthington]

Originally Posted by dwndrgn

The main reason that HAL (as an AI computer instead of just a 'regular' computer) was on the ship was to be the last resort crewmember - should anything happen to the biological crew, mr. mechanical crew would take over so that no matter what, images and information regarding the portion of the solar system the ship was travelling through and to would get back to earth. The ship itself was never expected to go back to Earth so the images and info had to continue to be collected and sent back.

Well, according to the film -- in some ways vastly different from the novel -- HAL 9000's was ostensibly there to take over in case something happened to the human crew, but in reality HAL was the only one given the full information on why the mission was put together in the first place, what the purpose of the mission was, and the "proper" response should the recipients of the signal from the monolith prove to be a threat (if they still existed). Even the members in cryonic suspension had specialized compartments of knowledge rather than the whole picture. And then they did the typical human thing of expecting a logical machine which would not be capable of understanding intuitively the requirement for such, TO LIE, WHILE APPEARING TO BE DEALING ENTIRELY HONESTLY, WITH ITS FELLOW CREW MEMBERS. While HAL has intelligence, and is programmed to imitate emotive response (to a degree, remember the flat affectless delivery), he is incapable of emotion as such, so that he is also incapable of truly understanding the "need" for secrecy, especially toward those he is to take care of and nurture -- but only as long as they do not threaten the mission by trying to find out things they do not already know.

So ... HAL was not there as backup, but actually as the primary member of the crew, because he could be trusted to not reveal that which he was programmed not to. Unfortunately, it would have failed in any case, because the human element was needed to make the intuitive decisions of such a delicate and important mission.

Also, unless I'm seriously misremembering, the Discovery was indeed intended to return to Earth, though never meant to land. There was no intend to abandon the astronauts, and no one foresaw the effects on HAL of the contradictory programming he was given. As for the removal of portions of HAL's "brain" being possible, while leaving the computer analog of the autonomic motor functions this is, again, given as a deliberate choice in computer design because of the unknown factors of such an AI; there is a hint that the mission was actually sped up long before they would have preferred to go that far, by the perceived danger of the signal from the monolith -- which Frank and Dave knew nothing about, if you'll recall, as the entire thing about the monolith had been kept very much secret. The other crew members knew about it, of course; and this was also part of why they were sent in hibernation units rather than allowed to interact freely with Frank and Dave. Poole and Bowman were the "drivers", and ultimately expendable if need be; the other crew members were not.

[Jim Colyer]

I read the book and saw the movie. I do not think either hold up particularly well.