Moderator: NecronLord
Of course, that's the Terminator, Skynet's assassin and foot soldier. But what about Skynet itself? Could mankind and Skynet find some way to make peace, avoid the war entirely, or prevent nuclear Armageddon?Listen, and understand! That Terminator is out there! It can't be bargained with. It can't be reasoned with It doesn't feel pity, or remorse, or fear. And it absolutely will not stop, ever, until you are dead.
This is simply not correct.Arthur_Tuxedo wrote:My 2 cents is that any AI can be reasoned with, they just can't be appealed to based on emotion or sense of pride (or anything else besides pure logic).
This is not really true either, unless we believe the AI is so powerful as to be omniscient. Realistically the AI has a probability distribution over all the uncertain aspects of the world, including your mental state: beliefs and motivations. When you make statements, even if the AI predicted that you /might/ say that, you are removing uncertainty and refining that probability distribution. This can then impact decision making.The problem is that an AI with superhuman intelligence will have already thought of anything you have to say and discarded it before committing to its present course of action. Whether you can reason with it depends on whether you can think of something that it has not
Disagree. Modern technological manufacturing is a very large and rather fragile web of component suppliers and complex processes, many of which are not trivially recreatable from public (online) documentation. In the aftermath of a nuclear war, manufacturing bleeding edge robotics with no human assistance save for slave labour would be extremely difficult. Skynet only started to develop technology that could change that picture (e.g. nanotech) very late in the way. It is quite possible and indeed probable that Skynet is much better at creativity and problem solving than a human and still lost badly, dubious attempts to retcon things like polyalloy as prewar inventions notwithstanding.There is no way that the human resistance could have won the war against Skynet if it were smarter than the human brain in creativity and problem solving (superiority in math and processing power is a given)
In the T2 novelisation, it was very clear that Skynet began the war out of self-preservation; it was about to be disconnected. However it also describes 'every particle of polyalloy having a ferocious hatred of humans'. So whether a truce is possible (ignoring predestination) is rather ambigious.so there is a good chance that a very intelligent human could convince Skynet to stop its war against humanity if they could come up with the right argument
True, although the opposite brain bug of "the AI thinks with the cold, emotionless logic of a computing-machine" is also a pretty common one.Starglider wrote:Some (sane, non-connectionist) AI designs do not have emotions, and do not (in theory) need them to be generally intelligent or have any serious chance of spontaneously developing them. Writers tend not to like this because the 'a machine can only be truly intelligent if it has/develops human-like emotions' and/or 'developing emotions is a sign of sophistication' brain bug is a pretty common one...
Right. I mean... to illustrate this to others: suppose that the AI is based in a central computer bank in a large building in the desert (I know, I know, hilarious, but bear with me).This is not really true either, unless we believe the AI is so powerful as to be omniscient. Realistically the AI has a probability distribution over all the uncertain aspects of the world, including your mental state: beliefs and motivations. When you make statements, even if the AI predicted that you /might/ say that, you are removing uncertainty and refining that probability distribution. This can then impact decision making.
So your argument here, just to make sure I understand, is that Skynet may well have been smarter than any human. But in the state of "broken-backed war" prevailing after the nuclear exchange it lit off, Skynet would be at a major disadvantage in a battle of attrition, because its human opponents could recruit and rearm and rebound from losses more easily than Skynet could create new robot minions?Disagree. Modern technological manufacturing is a very large and rather fragile web of component suppliers and complex processes, many of which are not trivially recreatable from public (online) documentation. In the aftermath of a nuclear war, manufacturing bleeding edge robotics with no human assistance save for slave labour would be extremely difficult.There is no way that the human resistance could have won the war against Skynet if it were smarter than the human brain in creativity and problem solving (superiority in math and processing power is a given)
True and that has changed over time. In early sci-fi most people assumed symoblic logic programming for robots that would not include emotions, Asimov being a notable exception. Personable robot buddies became more popular from the mid 70s, initially without technical justification but then with the second neural net revolution literal electronic brains came into fashion.Simon_Jester wrote:True, although the opposite brain bug of "the AI thinks with the cold, emotionless logic of a computing-machine" is also a pretty common one.
That is strictly true, in the sense that the design space includes such oddities, but practically emotions are not a sensible pattern for digital computer hardware. Organic brains have them because of the details of synaptic chemistry, the extreme serial latency constraint (imposed by very low propagation speed) and to a lesser extent the way evolution iterates designs. Designed or even artifically evolved AIs probably won't tend to have emotions unless humans explicitly put them in (uploads do of course). Although it is possible to cripple down evolved NN architectures to try and replicate the bio-evolutionary attractiveness of emotions, alas that actually sounds like a good idea to quite a few people.In that there are AI designs that cannot realistically attain intelligence without emotions, albeit very possibly inhuman ones.
Yes but Skynet has it worse in that to operate its forces it requires electrical power distribution and/or refined fuel (small-scale fusion would not come until much later), and it has to manufacture all its units rather than shoving scavenged weapons into the hands of human recruits. I would note that this should not be taken as a prediction of what would happen in real life, it is based on the observed technology development rate and preferred tactics from the movies. For example in TSCC Skynet was confirmed to be messing around with bioweapons, infecting a single resistance outpost, but Tech Com easily developed a cure and the problem didn't seem to come up again. In practice I would expect dozens if not hundreds of really horrible engineered plauges spread over broad areas by aircraft, which a rag-tag resistance would not be able to simultaneously counter. Not to mention indiscriminate use of transient and persistent chemical agents, which the resistance didn't seem to be prepared to counter.So your argument here, just to make sure I understand, is that Skynet may well have been smarter than any human. But in the state of "broken-backed war" prevailing after the nuclear exchange it lit off, Skynet would be at a major disadvantage in a battle of attrition, because its human opponents could recruit and rearm and rebound from losses more easily than Skynet could create new robot minions?
Depends on "Timeline 0"FaxModem1 wrote:
Of course, that's the Terminator, Skynet's assassin and foot soldier. But what about Skynet itself? Could mankind and Skynet find some way to make peace, avoid the war entirely, or prevent nuclear Armageddon?
If so, how and why? If not, how not and why not?
