|
|
Post by csj on Apr 28, 2017 6:52:03 GMT
But, if Robot does switch to a new head (and presumably a new processor) - would he cease to exist? Take that, robotheists. Perhaps. After all, the cognition, thoughts, feelings and perspective of *snuffle* are likely to differ between her two forms. And then, there's the question of if the totem switcheroo is flawless, whether the new body is in fact blank (I'm sure the Court would be tempted to mess with the process), etc. Less tenuous than real-world philosophical questions on the matter due to the 'soul theory' being scientifically-provable (or at least, testable) in this setting. |
|
|
|
Post by warrl on Apr 28, 2017 16:07:46 GMT
[being scientifically-provable (or at least, testable) in this setting. Strictly speaking, in the hard sciences nothing is provable. For a theory to be accepted, it must first explain all relevant pre-existing evidence; second, in some specific circumstance it must predict a different outcome than the previous theory; third, such a specific circumstance must happen and the theory's prediction must be correct - or, at least, closer to correct than the previous theory. However, there is ALWAYS some specific circumstance that is significantly different enough from what we've seen that the theory COULD BE wrong. What we have is a stack of theories that, if they are wrong, they are only wrong in what we would regard as extreme or bizarre circumstances. Example: Newtonian gravity is wrong - but only with very large masses, very high velocities, or very precise calculations over an extended period (e.g. GPS satellites, where the time it takes light to travel a meter is important). Einsteinian gravity seems to work under those conditions. Certain terms in Einsteinian gravity, when you're working with small masses and low velocities and don't care about extreme precision, get to be so small as to be insignificant - and if you drop them out, you get Newtonian gravity. |
|
|
|
Post by TBeholder on Apr 28, 2017 20:41:41 GMT
As long as the processor (well, in this comic at least. in reality it should've been the memory chip) is the same, wouldn't it be the same robot? There's more to it. First, those are not necessarily two completely separable things. And in a [quasi-] neural system there's not necessary only one long-term memory, either. Even in pocket calculator grade processors, there are at least von Neumann and Harvard architectures. A living brain is somewhat Harvard-like, in that it has separate "instruction" (trained net) and "data" (log) memory. We have seen that S13 running as an overdrive on disabled S1 had pre-set procedures, but no plain accessible knowledge - knows how to act, but not why. Which suggest they resemble a living brain in this. Log memory may also be individual, in that we don't even know whether any of them but "witness model" can record undigested input stream. There may be reasons to make the processing net is a part of CPU (who knows how hardware dependent it is, or it may be just stuck on the thickest bus for speed) or maybe it's brain-like enough that there's not much more to "CPU" other than this and interface. Then log memory may be thrown in (on the new models), whether also for speed or simply because it's not a bottleneck any more and "all in one" approach is convenient. |
|
|
|
Post by TheClockworkCoyote on Apr 29, 2017 0:55:53 GMT
Kat's clearly been skipping out on her mad science classes. There's supposed to be more maniacal cackling, challenging/mocking the gods, ranting about how the peasants will never understand your work, and use of lightning storms as backgrounds for scenes like this. I mean, yes, it's just you and Annie here, but that's no reason to get sloppy. Have some professional pride, dammit!  I think a particularly interesting part of Kat's personality is her belief that she is just a normal person. Creating Robot's body is just something a normal friend does if she knows how. No call for manical laughter if you're just making tea. I also expect Kat to have reasons to continue believing that she is just a normal person when others try to tell her she is the Mecha Angel. - Uh, robots just act funny sometimes. - Wah? Paz, you were still experiencing the effects of being in Zimmingham when you saw something in the blue smoke. - Come on Annie, we've seen things differently before and I could see my normal self. Further proof she's been skipping classes if she's incapable of recognizing such an opportunity to monologue in the traditional manner |
|
|
|
Post by ariestinak on Apr 29, 2017 3:10:44 GMT
As long as the processor (well, in this comic at least. in reality it should've been the memory chip) is the same, wouldn't it be the same robot? There's more to it. First, those are not necessarily two completely separable things. And in a [quasi-] neural system there's not necessary only one long-term memory, either. Even in pocket calculator grade processors, there are at least von Neumann and Harvard architectures. A living brain is somewhat Harvard-like, in that it has separate "instruction" (trained net) and "data" (log) memory. We have seen that S13 running as an overdrive on disabled S1 had pre-set procedures, but no plain accessible knowledge - knows how to act, but not why. Which suggest they resemble a living brain in this. Log memory may also be individual, in that we don't even know whether any of them but "witness model" can record undigested input stream. There may be reasons to make the processing net is a part of CPU (who knows how hardware dependent it is, or it may be just stuck on the thickest bus for speed) or maybe it's brain-like enough that there's not much more to "CPU" other than this and interface. Then log memory may be thrown in (on the new models), whether also for speed or simply because it's not a bottleneck any more and "all in one" approach is convenient. Somehow true. But this is based on the assumption that Tom knows enough about neurology, artificial intelligence and computer hardware to intentionally use the terminology "processor". Although it's important to know that robot's "brain" isn't reset after being dismantled (and thus, being plugged out of power for a while). So it definitely operates on a non-volatile memory (as opposed to all kinds of volatile memories that are commonly used inside the processors and get reset every time the power is cut). This is quite hard to implement in tiny processors and that's why the non-volatile memory in our computers are all (relatively) large peripherals outside the main processor. The other important note is that a neural net is both a processor and a memory. There is no separate log in our brains. So while it is theoretically possible for Robot to have a "neural net with on-chip non-volatile memory" as a processor, this configuration isn't practical or efficient due to the limits of hardware. |
|
|
|
Post by warrl on Apr 29, 2017 4:53:24 GMT
Non-volatile memories are all relatively large? Have you ever looked at a microSD card? Also, there's been some progress in putting many processors - each with some of their own memory - on a single chip. I wouldn't class it as a neural net... yet. |
|
|
|
Post by ariestinak on Apr 30, 2017 4:27:00 GMT
Non-volatile memories are all relatively large? Have you ever looked at a microSD card? Also, there's been some progress in putting many processors - each with some of their own memory - on a single chip. I wouldn't class it as a neural net... yet. Relative to the CPU-cores themselves, yes. And I said that it is possible to create a neural-net processor with on-chip non-volatile memory. The problem is that such a configuration is inefficient and impractical. We almost always put volatile SRAM memory on processors because it's simply cheaper and more practical than other options (And I guess it's more resistant to high CPU temperatures, being made from the very same stuff). There are actually neural-net processors being developed out there. They're just using volatile memory and accepting the loss of learned info after system reset (which is actually not so dissimilar to our own brain). |
|
|
|
Post by jda on May 2, 2017 14:40:26 GMT
so, on the topic of robot getting a new body and if it will still be robot, i'm just gonna leave these here; link linkInterestingly enough, there are 4 words used to describe the *it* that gets transfered on a body migration: Coyote: the mind, the will and the life Court: the spirit. Seems that Coyote can discern (as in separate) them one of another, while the Court only knows it as a pack... Or maybe (as implied before) can cheat and modify them (plant a bug on them?) on the Ritual for transfer. |
|
|
|
Post by jda on May 2, 2017 14:48:08 GMT
Well, reading here and here makes me wonder: Are the Fairies body born on the Forest AND their Totem (Mind/Will/Life) on the Court? So, they have to get killed for their totem to be free for transfering a human body. That would be a Hell of a HiSpeed ample range Wireless Connection.  |
|
|
|
Post by puntosmx on May 3, 2017 2:58:22 GMT
Somehow true. But this is based on the assumption that Tom knows enough about neurology, artificial intelligence and computer hardware to intentionally use the terminology "processor". Although it's important to know that robot's "brain" isn't reset after being dismantled (and thus, being plugged out of power for a while). So it definitely operates on a non-volatile memory (as opposed to all kinds of volatile memories that are commonly used inside the processors and get reset every time the power is cut). This is quite hard to implement in tiny processors and that's why the non-volatile memory in our computers are all (relatively) large peripherals outside the main processor. The other important note is that a neural net is both a processor and a memory. There is no separate log in our brains. So while it is theoretically possible for Robot to have a "neural net with on-chip non-volatile memory" as a processor, this configuration isn't practical or efficient due to the limits of hardware. Why not include everything in a single "pack? CPU, volatile memory (cache/RAM) for processes, firmware for OS/IO and non-volatile memory for memories? Also, we know that the cube is a 3d representation of a part of the code for a robot's system. Maybe consider it just the kernell. This all compressed into the chips that we've seen Robo be stored at..... those SoCs are awesome indeed. |
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on May 3, 2017 13:37:27 GMT
Somehow true. But this is based on the assumption that Tom knows enough about neurology, artificial intelligence and computer hardware to intentionally use the terminology "processor". He finished a degree in computer science. |
|
|
|
Post by Jelly Jellybean on May 3, 2017 16:58:42 GMT
Somehow true. But this is based on the assumption that Tom knows enough about neurology, artificial intelligence and computer hardware to intentionally use the terminology "processor". He finished a degree in computer science. He also plays the banjo and that is all the confirmation I need! 😜 |
|
|
|
Post by tc on May 4, 2017 6:52:19 GMT
Maybe the processor stores memory into the ether. It's Diego's work after all... I was pretty sure Robot was a later model (i.e. machine-designed and simplified) Seraph. As far as we know, Kat hasn't grok'ed what that "extra" bit on his personality chip is as yet, but it's implied to be rather less complicated than the "hearts" of Diego's original machines, no? As far as his head goes, it looks like there's an access panel on the lower-right (when facing him) that's currently removed - presumably allowing access to the chip and in this case providing an I/O interface to Kat's test rig. |
|
|
|
Post by keef on May 4, 2017 8:06:10 GMT
Maybe the processor stores memory into the ether. It's Diego's work after all... I was pretty sure Robot was a later model (i.e. machine-designed and simplified) Seraph. True. But we don't know how old his CPU is. As it fits the socket in S1's head, the CPU model must be around since Diego's days. ( Intel and AMD could learn something here... ) I always thought it had something to do with Robot's prophesies. Anyway, when Renard's personality seems to change the exact moment he changes shape, robot's mind seems to be more than a plaything for his body. Although his experience of pain and his love for Shadow may be changing him now. |
|
|
|
Post by tc on May 4, 2017 10:02:29 GMT
True. But we don't know how old his CPU is. As it fits the socket in S1's head, the CPU model must be around since Diego's days. ( Intel and AMD could learn something here... ) Come to think of it, I wonder if art is actually imitating life to some extent here. Bear with me...  For starters, I should have been a bit more clear on the term "simplified" - in tech terms, it doesn't necessarily mean that each component becomes more basic; sometimes functions requiring multiple components in earlier designs are merged into one component in later designs if there's a good reason for doing so (e.g. adding basic I/O functions to the 6502 to make the 6510 used in the Commodore 64, reducing the chip count and overall unit cost), thus simplifying the overall design. We know Diego's original robots had "hearts" which seemed to serve both as an etheric power source and possibly more complex (emotional?) functions, and we know that after Diego died, his robots set about designing simplified versions of themselves based on the aspects of themselves that they *did* understand. We've also seen that the "modern" robots (aside from Robot himself) appear to have much more basic "personality" and behavioural subtlety than Diego's. So my wild-arsed guess would be something along the lines of the following : The chip which fit the socket in S1 (Diego's design) was nothing more than a physical logic controller, with the "mind" and memory aspects stored and handled elsewhere. This would explain why, when S13's chip was installed in S1, S13's "mind" was overridden by those other components within S1. WAG 2: If the robot "personalities" are indeed more simple, yet still etheric; I have to wonder : Robot didn't show signs of having exceptional emotional capacity until after he returned across the bridge under the control of "false" Shadow, who tried to harm Annie (though this may only be because Annie was the first person from the Court encountered on the bridge). Robot is aware that the seed Ysengrin dropped in his arm socket (presumably at Coyote's behest) grew to form a wooden arm, but we don't know if that seed changed other things... After all, if Robot's "mind" is etheric - and Coyote's etheric abilities are effectively unlimited - it's well within the realms of possibility that Coyote's seed could have expanded the capabilities of Robot's etheric "mind", intentionally or otherwise. |
|
|
|
Post by keef on May 4, 2017 14:22:59 GMT
For starters, I should have been a bit more clear on the term "simplified" - in tech terms, it doesn't necessarily mean that each component becomes more basic; sometimes functions requiring multiple components in earlier designs are merged into one component in later designs if there's a good reason for doing so ( - ) thus simplifying the overall design. We know Diego's original robots had "hearts" which seemed to serve both as an etheric power source and possibly more complex (emotional?) functions, and we know that after Diego died, his robots set about designing simplified versions of themselves based on the aspects of themselves that they *did* understand. We've also seen that the "modern" robots (aside from Robot himself) appear to have much more basic "personality" and behavioural subtlety than Diego's. Yes, Frankenbot, for example, comes over as much more mature than the modern court robots. No, Annie was holding Renard, and the shadowman's mission was to retrieve Renard from the court. Annie was just in the way. We don't know, we do know Robot was impressed, possibly changed, by the fact Annie gave him a choice. ( and Robot had been dismantled earlier because of his prophecies about the coming of the Angel. He was always special.) We have a Wild-arsed Speculation thread here you may like it. |
|
