Comic Discussion > QUESTIONABLE CONTENT
WCDT Strips 3461-3465 (17-21 April 2017)
Thrudd:
--- Quote from: JoeCovenant on 25 Apr 2017, 02:32 ---
--- Quote from: Thrudd on 24 Apr 2017, 11:56 ---The absence of evidence that something does not exist is not evidence that it exists.
--- End quote ---
You're absolutely right.
But this is a made-up universe, upon which we can only transfer our 'own world' realities, until we are presented with something that actively stands contrary to what we know.
What we know is, computers store data which they are fed, and unless pre-programmed to do so, will not lose that data...
And logically, what sense does it make to program a Combat AI to have the capacity to *forget* what it's squad mates look like?
(Mind you, that might deny the old oxymoron... "Military Intelligence" :) )
There is no evidence that memory loss *does* happen, so that's all we can go with.
--- End quote ---
In theory but I occasionally have to work with computerized control systems as well as ye old relay logic control interlocks.
Guess what, they do loose data. Sometimes it is just a bad memory cell on a control module or media degradation or just a stray cosmic ray flipping a bit or two [Murphy at play].
So there is that.
Then there is assumptions in general - you know what that word implies right?
A good analyst looks at the evidence first before building a model, not the other way around.
And again, I reiterate, that the absence of evidence that something does not exist is not evidence that it exists.
So with respect to the memory loss and the volatility of memories in an AI we have at the moment exactly ONE data point.
(click to show/hide)Insert cheeky comments regarding inductive reasoning and creativity
One gets such wholesale returns of conjecture out of such a trifling investment of fact. - TWAIN
or I can just throw in an xkdc reference :-D
JoeCovenant:
--- Quote from: Thrudd on 25 Apr 2017, 07:03 ---In theory...
--- End quote ---
Theory is all this is working on.
--- Quote from: Thrudd on 25 Apr 2017, 07:03 ---but I occasionally have to work with computerized control systems as well as ye old relay logic control interlocks.
Guess what, they do loose data. Sometimes it is just a bad memory cell on a control module or media degradation or just a stray cosmic ray flipping a bit or two [Murphy at play].
So there is that.
--- End quote ---
Which I have stated more than once now, I not only accept, but agree with.
--- Quote from: Thrudd on 25 Apr 2017, 07:03 ---Then there is assumptions in general - you know what that word implies right?
--- End quote ---
Is this "insult the newb" day , or what??? I find that comment quite insulting.
--- Quote from: Thrudd on 25 Apr 2017, 07:03 ---A good analyst looks at the evidence first before building a model, not the other way around.
And again, I reiterate, that the absence of evidence that something does not exist is not evidence that it exists.
--- End quote ---
I think you might mean "absence of evidence that something does not exist, is not evidence that it DOES NOT exist"... no?
But that has exactly what to do with any comment I have made?
I have not argued that lack of evidence of anything.
I have argued the only evidence we have is real life versus comic reality... and comic reality has done nothing to counter that.
But until such time as the issue is actually addressed 'in comic', what else have we to draw from?
I would imagine this is why the rules of 'canon' exist... and an absence of evidence that something exists, does not make anything enter comic canon.
But... one last time... (as I am beginning to get annoyed at the casual swipes at my intelligence).
My concern is simply that Bubbles entire trauma did not need to be 'bolstered' by the (to my mind) silly statement that she cannot remember the faces of her squad.
I DEEPLY hope it comes out that those very memories have been deliberately purged by someone... otherwise it is simply not believable. The suggestion that ANY person, AI or otherwise, would so easily forget the faces of their colleagues... ESPECIALLY those they were in combat with... flies in the face of 'real-life' experience.
I shan't go into too much personal detail here, but I had a great deal of experience when I was younger, discussing 'Living History' matters with WWII veterans. All of them at some point said the same thing. They never forgot those experiences, and can recall their 'lads' as if it were yesterday. My grandfather was one of those, and he was shot in the face (and survived, thankfully).
Even on a personal level. my memory is hardly photographic, but I can plainly see the faces of almost all the kids in my primary seven class... that's over 40 years ago.
If an AI can't retain memories longer than that, for people who are particularly important to them (particularly in the military), is that an acceptable tolerance?
(Particularly when such persons have been created to feel emotions.)
--- Quote from: Thrudd on 25 Apr 2017, 07:03 ---So with respect to the memory loss and the volatility of memories in an AI we have at the moment exactly ONE data point.
--- End quote ---
Yes... and that is the VERY CRUX of the problem I have with it.
That ONE data point came out of nowhere and counters any physical and intuitive proofs we have IRL.
As the youngsters say these days... End. Of.
sitnspin:
--- Quote from: JoeCovenant on 25 Apr 2017, 02:32 ---And logically, what sense does it make to program a Combat AI to have the capacity to *forget* what it's squad mates look like?
--- End quote ---
AI in the QCverse are not programmed to do anything. They are emergent, same as biological intelligences. Bubbles chose a career in the military. She was not designed to be a soldier, it was a deliberate personal choice on her part.
Besides, it's already been heavily implied that her memory of them was erased, not that she just forgot. That's part of the tragedy of the situation. She had her memory of them and the circumstances around the trauma she wanted to forget erased, but she's still left with the emotional impact of those memories.
blt:
--- Quote from: sitnspin on 25 Apr 2017, 12:30 ---
--- Quote from: JoeCovenant on 25 Apr 2017, 02:32 ---And logically, what sense does it make to program a Combat AI to have the capacity to *forget* what it's squad mates look like?
--- End quote ---
AI in the QCverse are not programmed to do anything. They are emergent, same as biological intelligences.
--- End quote ---
This, and Jeph's comment that they are born and develop in a "creche" until they bootstrap themselves into full intelligence, should really be all there is to say on this debate. They don't behave like any real world analogue and the answer is Sci-fi handwaving.
--- Quote ---I DEEPLY hope it comes out that those very memories have been deliberately purged by someone...
--- End quote ---
Bubbles asks.
CW provides.
In a way.
Tova:
Hey Joe, I am sorry about the snark. I can act like a bit of a dickhead sometimes.
I didn't mean to take a swipe at your intelligence though. Honestly, I was actually shooting for 'stubborn bastard.' ... I hope you're laughing now, or I'm really in the shit.
On a more serious note...
--- Quote from: JoeCovenant on 25 Apr 2017, 08:46 ---I DEEPLY hope it comes out that those very memories have been deliberately purged by someone... otherwise it is simply not believable. The suggestion that ANY person, AI or otherwise, would so easily forget the faces of their colleagues... ESPECIALLY those they were in combat with... flies in the face of 'real-life' experience.
I shan't go into too much personal detail here, but I had a great deal of experience when I was younger, discussing 'Living History' matters with WWII veterans. All of them at some point said the same thing. They never forgot those experiences, and can recall their 'lads' as if it were yesterday. My grandfather was one of those, and he was shot in the face (and survived, thankfully).
--- End quote ---
Good point. On the general topic of total recall we may still differ (though I'm not sure one way or the other, just leaning), but here I believe you are 100% correct.
I suspect, though, that the answer is that this was collateral damage to the purging that CW carried out.
Just to reiterate my thoughts on AIs and total recall... I don't believe that AIs 'forgetting' something in the general case means that they have volatile memory necessarily - although they might. But it's more likely that it means that the memory temporarily cannot be brought into the conscious mind.
I think that there probably are real-world analogies for this in current AI and non-AI software, and thus I don't think that an AI without total recall would contradict what we currently know about computers. As an example in the realm of AIs, google 'catastrophic forgetting'. I'm not at all saying that this is the precise mechanism that might be at play in QC AIs, merely that it is conceivable that an AI might forget something.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version