[News] AI will be smarter than humans in 30 years

[Sparky]

@MattiasNYC:

You're saying we could create something smarter than ourselves, basically. How do you really, logically, come to that conclusion? How is that logically possible? How would a lower order being create a higher order being, in essence?

Create something that is faster than us? Sure. But smarter? As in, create and develop stuff, thought processes, etc, beyond what we can comprehend and grasp? Really, really, really doubtful. I'll grant you the 0.1% chance that maybe that is possible, I guess. But I guess you come to this conclusion because you must obviously believe that we all originally came from some sort of primordial ooze and life, thought, etc, is all purely an "accident" and has no real meaning, purpose, or end. I guess coming from that worldview, yeah, sure, it can happen. But... that worldview introduces far too many more problems than it "solves" - and the one core problem it attempts to solve (or rather bypass) it still doesn't satisfy, ultimately.

The end problem with this whole topic though is it does start necessitating delving into stuff that really this forum doesn't permit, so, I'll just end it there.

Suffice to say, while we could create an "AI" that is pretty darn smart and fast, it still won't surpass the human race as a whole. Sure, we could do some serious damage to ourselves with it if we're not careful And it wouldn't surprise me if we end up doing that... we can be pretty darn stupid playing with stuff sometimes!

[MattiasNYC]

You are making a bold a affirmation that you cannot prove. You can assume that there is no God just like I assume there is one.

That was exactly my point.

The denial of existence of God is just as religious as the recognition of the evidence, since neither of them can be proven.

By definition; no, it is not. Atheism is by definition NOT theism. Saying it is is just nonsense.

My second point is that all those articles stand on assumptions which in turn stand on other assumptions.And on top of those assumptions we have news that sell. Just changing the root assumption from evolved universe to created one changes the possible outcome for humans ever being able to generate a higher intelligence.

Not necessarily. You would have to show how that is the case. If I was to accept for the sake of discussion that god created man, there is absolutely nothing there that prevents man from creating something more intelligent than man.

And to add to the topic, might be better to define what kind of intelligence can and cannot be developed and what could count as true intelligence and what not. We are already at the point where there is almost no game where humans can beat the computers, however this is pure brute force, no matter the algorithm used.

In exactly what physical sense is human thought not a matter of "brute force"? If you have one neural network that happens to be "natural" and biological, and you have one that is "artificial" and man-made, in just what sense is there a difference apart from how they were created?

There are also other ways to validate an intelligence like putting it to come up with rational provable explanations to existential questions. And we could start with:
- how was the first star formed? (you have to first compress H2 and He2 to some point before gravity can take place)
- how were the first planets formed? (gravity according to simulations cannot create large bodies by aggregation)
- how life came from non life? How information came into existence? and who designed the architecture? (DNA is like a source code in base 4, ~770MB/human genome, however in my experience, source code is useless if you do not design a computer architecture that can load and execute the source code).
- how many genetic mutations and how many generations are needed from first cell to Homo Sapiens Sapiens and how much time given the minimum generational age between mutations (mathematical models for evolution show it's possible only when entity is very small and generation time is smaller than 3 months).
And I could add some bonus questions, like why there is a constant layer of marine fossils all over the world like somehow it was covered in water all at once, why all fossils are actually result of rapid burial, why there is C14 found consistently in diamonds when there is no physical way to contaminate them, why the diffusion rate of He2 in rocks found in the deep of the earth suggests accelerated decay and an earth age of about 6000 years? And the list could go on.

Are you sure you want to argue that we can validate intelligence by saying that the earth is 6000 years old?

Now a few observations from software development:
- there is at least a bug for every 10-100 lines of code (close to lower rate)
- subtle bugs take years to be discovered and sometimes complete redesign for fixing
- concurrent programming is hard and concurrency bugs are transient many times and hard to fix
- the quality of the developers that graduate is worse and worse every year

I think the above illustrates quite well the problem. You are entirely ego-/human-centric in your view on everything. And that is as is expected considering your religious views. So what I would argue you are missing is that just because you design certain software a certain way a) doesn't mean other people do it the same way, and b) most certainly doesn't mean that an actual AI would do it that way.

If you consider a human being for example you'd see that we're incredibly flawed animals, yet we're very resilient. Not necessarily roach-level resilience, but we can certainly adapt in amazing ways. Just about an hour ago I saw a video on a news website about a girl who suffered serious seizures repeatedly as a child. It was so bad the only solution was to literally cut out half of her brain, which they did. Now she is almost entirely normal, and there are zero signs that she's literally missing half a brain. So from the standpoint of your argument above, this human had a ton of bugs per X lines of code, bad enough to affect, or "infect", 50% of the computer (the brain). The solution in that biological neural network was to re-do the learning which was possible because of the brain's plasticity at the early stages of development. So, similarly to that, we simply don't know what a 'properly' developed AI is capable of. We can't assume that we as humans have to fix its bugs. We don't even know that the bugs need fixing in the first place. All of that line of reasoning appears to be 'linear', as was pointed out before. You write code. Code goes from one place to another in a predictable way. Therefore that must always be the case. But that's exactly the point people are making; that doesn't necessarily have to be the case in an AI.

Given the observations, you may understand why I (and I think many other engineers) do not believe it will ever be possible to develop something that can communicate in an intelligent manner with humans and be able to contribute to the society by answering existential questions. However, if I would live to see it, I would not be surprised if the AI would answer to all the existential questions with one world: God.

I'm sure that an AI very well might tell you that god was behind it all, but not necessarily because it reached that conclusion. In fact, there could be huge amount of reasons for it telling you that, most of which would not be based on its belief in a god. In other words it could lie to you, and you would believe it because it affirmed your preconceived notion. Heck, you've even pretty much just set up a fail-safe scenario: You are convinced that god exists and are claiming that the answer to whether or not there is intelligence is if that answer can be reached, that god exists. So if the AI tells you that god exists - if that is the conclusion it reaches - then you feel that it is intelligent. If it comes to the conclusion that god does not exist then by (your) definition it is not intelligent, and you can ignore it's conclusion. Win win.

But much worse: Suppose it gains a tremendously higher intelligence - would you even be capable of communicating intelligently with it? When was the last time a human had any deeper meaningful conversation even with the next smartest species on the planet?

People have to stop thinking about this linearly, and think about it non-linearly.

[MattiasNYC]

@MattiasNYC:

You're saying we could create something smarter than ourselves, basically. How do you really, logically, come to that conclusion? How is that logically possible? How would a lower order being create a higher order being, in essence?

I think the question should be why it wouldn't be possible.

Create something that is faster than us? Sure. But smarter? As in, create and develop stuff, thought processes, etc, beyond what we can comprehend and grasp? Really, really, really doubtful. I'll grant you the 0.1% chance that maybe that is possible, I guess.

I actually think "speed" is relevant in how we evaluate intelligence. And I do think that this again is all a "non-linear" issue. Obviously if the way you think about it is that a computer would have to be programmed not only for it to understand how to do things but what to do then of course there seems to be a limit; the limit is [u]us[/i]! And that's why I keep saying the argument is "linear" and ego-centric. But if an AI actually gained consciousness, even in a very primitive limited sense, it could actually create its own goals. This is like a "switch" that gets flipped which is why it isn't at all necessarily "linear".

But I guess you come to this conclusion because you must obviously believe that we all originally came from some sort of primordial ooze and life, thought, etc, is all purely an "accident" and has no real meaning, purpose, or end. I guess coming from that worldview, yeah, sure, it can happen. But... that worldview introduces far too many more problems than it "solves" - and the one core problem it attempts to solve (or rather bypass) it still doesn't satisfy, ultimately.

But your dissatisfaction with not knowing all the answers doesn't mean there is a god. That's just an "argument from ignorance". In other words wherever your knowledge ends god begins. It's an illogical though process, and history shows it.

Further more, I don't think us creating an AI is more likely possible because there is no god, I just think it's possible and not unlikely even if there is a god. There's nothing in religious views that tells me it wouldn't be possible because of god's existence. Really the only religious argument against it is that it would make god less special, and by extension humans as well.

The end problem with this whole topic though is it does start necessitating delving into stuff that really this forum doesn't permit, so, I'll just end it there.

That's just because you took it there. There is absolutely nothing technical or natural that relates to the topic of A.I. becoming more intelligent than humans that point us towards religions. Clearly you have to believe in god to get there, which is telling.

Suffice to say, while we could create an "AI" that is pretty darn smart and fast, it still won't surpass the human race as a whole. Sure, we could do some serious damage to ourselves with it if we're not careful And it wouldn't surprise me if we end up doing that... we can be pretty darn stupid playing with stuff sometimes!

Again though I wouldn't be so dismissive of even that possibility. Currently the amount of computational power that exists even in my cellphone is absolutely stunning when put in a historical context. And I can connect with someone on the other side of the planet in mere seconds, and we can exchange information with each other close to instantaneously.

Now, fast forward 30 years. Where would we be in terms of computational power? Where would we be in terms of interconnectedness? Just how do we know that this future AI doesn't use computational resources all over the planet if it has the option and chooses to? Because we'd know and would pull the plug? Are we sure we'd know?

[Sparky]

You made a few assumptions about "where knowledge ends there god begins" - no, I don't believe that at all. I do agree that is illogical. But you assumed I'm thinking this, then took that running to a wrong conclusion based on a bad assumption

Define this linear/non-linear thing. You keep referring to it, and maybe it is because I'm up late working and my brain is tired, but it isn't quite clear to me right now exactly what you're referring to. Doesn't help you keep putting it in quotes, which makes me think it is being used perhaps loosely?

[MattiasNYC]

Yeah, sorry about assuming something that wasn't the case. To me it looked as if you were basically saying that since we can't explain certain things using the scientific method it's likely that god was the cause. If that wasn't what you meant to say then I misunderstood you.

"linear" in this case is indeed a bit loosely used. I'm tired as well so I doubt I can explain it well right now. Maybe tomorrow. You can maybe read Particle's post (7) because I think he puts it succinctly.

[kromosto]

It is not possible to make a real AI (even it doesn't have to be smarter than human) with digital computing. You can simulate an AI, you can make it efficient and fast on some fields than a human and call it that it is smart. Etc. etc. But being digital is the limit. If you know quantum theory you can say on quantum level we may count as something like digital too but level is very different.

Now there is quantum computing. It can have chance but again the problem is programmers still think digital. First it have to change. 30 years is a long time. Humans are getting smarter so while I am not closing the door totally but making a real quantum computer and making a revolution on our brains about how we program machine is not seem possible to me. Yes we will see things that will amaze us on AI and deep learning in 30 years... but real AI. That is the problem.

[sergiu]

@MattiasNYC
The negation of a religious position without evidence that does not require the use of faith is still religion. No matter how you define it in dictionary.

You can call my view ego/human centric and I'm fine with it. It does however not change the facts. Writing code is hard and bugs grow exponentially with number of lines of code. Added complexity means design bugs. Usually bugs corrupt the functionality, do not enhance it or add new features. Given the software complexity, one must have a huge amount of faith to believe we humans can write something of such high complexity that can achieve true levels of intelligence. Reductionism tends to ignore the details of the implementation.

As I said, I gave one of many possible ways to evaluate intelligence by challenging such an AI to come up with reasonable explanations for existential questions. The set of questions listed in my previous post have until now explanations that require faith (I must have faith for example that radioactive decay rate stayed constant while there is scientific evidence that points to periods of accelerated decay). I'm willing to bet an AMD ReRyzen (AMD's CPU available after 30 years) that no AI will be able to pass such a test.

Offtopic: My religious views are backed by science. Science goal is to find the truth, no matter where it leads. Taking a position and enforcing it even when there is evidence against it's at best bad science and at worst religious dictatorship.

[MattiasNYC]

@MattiasNYC
The negation of a religious position without evidence that does not require the use of faith is still religion. No matter how you define it in dictionary.

Ok, so words really have no meaning

You can call my view ego/human centric and I'm fine with it. It does however not change the facts. Writing code is hard and bugs grow exponentially with number of lines of code.

Sorry, but that doesn't sound like a fact at all to me. Exponential growth relative to lines of code? Everything I've read and seen tells me that not only is the amount of bugs per line of code relatively similar in popular programming languages, but also that it scales more or less linearly relative to the amount of lines, not exponentially.

Added complexity means design bugs. Usually bugs corrupt the functionality, do not enhance it or add new features. Given the software complexity, one must have a huge amount of faith to believe we humans can write something of such high complexity that can achieve true levels of intelligence. Reductionism tends to ignore the details of the implementation.

But again, the above is an example of ego-centrism and a "linear" mindset. You assume that because humans are currently programming computers the errors that occur can only be solved by humans. And then you apply your perception of our limitations to that process.

But what people are afraid of isn't just that humans somehow manage to code AI better, but that we get AI that can modify its own code. Once that AI reaches a sufficient proficiency at doing so it'll do so faster than we can, and at that point is just what the limits of its performance really would be. If it takes you one day to find and fix two bugs in X lines of code, and it takes the AI one hour, then once it's improved itself by fixing its own code, how do you know it won't fix the next two bugs 5% faster? And so on.

Your limitation as a human being isn't automatically for all future limiting AI. You don't know that that's true. That's the point. This isn't necessarily scaling linearly.

As I said, I gave one of many possible ways to evaluate intelligence by challenging such an AI to come up with reasonable explanations for existential questions. The set of questions listed in my previous post have until now explanations that require faith (I must have faith for example that radioactive decay rate stayed constant while there is scientific evidence that points to periods of accelerated decay). I'm willing to bet an AMD ReRyzen (AMD's CPU available after 30 years) that no AI will be able to pass such a test.

Your problem above is that you have taken something that you are essentially saying can't be proven, assumed an answer based on your faith, and then said that the AI must be able to answer the question the way you currently like it for it to be considered intelligent. If your own answers are unreasonable then of course an AI giving you a reasonable answer will be "wrong" in your opinion. So your argument is just not not logical.

Let me just illustrate the problem by pointing out that these "existential questions" that you're talking about apparently have one answer; 'god did it'. Now, you're using that answer to those questions "to evaluate intelligence". By extension, it is logical to assume that you view anyone that doesn't reach the same conclusion as unintelligent. If that is not the case then it is possible for an AI to not reach that conclusion (like the human) yet still be considered intelligent (like the human).

Do you see how your reasoning is both illogical and taking you down a path that's not that great?

Offtopic: My religious views are backed by science. Science goal is to find the truth, no matter where it leads. Taking a position and enforcing it even when there is evidence against it's at best bad science and at worst religious dictatorship.

Your religious views, if they are that the earth is only a few thousands of years old (rather than billions of years) are most certainly not backed by science. You might think they are, but they aren't. The very same science that gave us everything from large airplanes and spaceships to microwave ovens nukes and the very computer you're typing on is what doesn't agree with you.

Then again, there's nothing "linear" about faith...

[sergiu]

@MattiasNYC
You described the standard chicken and the egg problem. In order to have an AI that can fix its own bugs you might have to fix the unfixable bugs first.
Regarding the code complexity, I recommend you good books like "The Mythical Man-Month" by Fred Brooks. Estimated complexity increases exponentially not linearly with number of lines of code in one project. You should ask any software architect or humble developer. I'm stating this as a developer with years of experience, not by reading on wikipedia. And until now I have not yet encountered a software architect that is not aware of the increased complexity. The holy grail is code reusage by design which allows a non exponential increase of lines of code with added features, but this does not invalidate the original statement.

I think we can reason on the vast topics on private. I just gave you a hint: how do you explain the existence of C14 in diamonds that are supposed to be billions of years old?

[Ket]

The way things have gone lately Skynet is preferable for ruling the human race. Americans voted for Trump, after all. If there are people who really thought voting for him was a good idea, a man who gave a speech about Uranium (or was it Plutonium? I forget, I'm not American so don't pay much attention to the rambling of a foreign madman) and clearly had no idea what he was talking about yet tried to act condescending to everyone in that speech, on top of essentially lumping nuclear weapons in to the "good" category while talking about Uranium then vaguely brushing off everything else nuclear as "bad", I'd rather have a machine dictate humanity any day

[sergiu]

@Ket
What you see is what happens when you leave out absolute moral standards. Good becomes evil and evil becomes good. Cannot however be fixed by a machine if the machine is programmed by the same people... It might however press the button faster.
There's another point that I've missed until now: while I do not entertain the idea that it's possible to have an AI more intelligent than humans, I see now the possibility that the human intelligence can be suppressed to a point where it's below or at the level of AI (no own will, no ability to discern between good and evil, no moral laws). Think about it: all media is about controlling how and what you think. Everything is designed to be "easy". Could it be that the outcome is the human kind inability to reason in 30 years?

[Sparky]

Given that skynet wanted to eradicate humans, I don't see how any reasonable person would want it in charge of humans

Don't let your extreme hate for a person blind you to everything, geez. I loath Hillary, she's the epitome of some of the worst political corruption, but I still wouldn't want Skynet over her.

Ouch.

[MattiasNYC]

@MattiasNYC
You described the standard chicken and the egg problem. In order to have an AI that can fix its own bugs you might have to fix the unfixable bugs first.

That's your assertion, and I see no reason that it is true. As I pointed out, even the most complex AI known to man, the human brain, could apparently have half of it plain cut out and the rest adapted. So it's really a question of plasticity. (and that ignores the fact that you call some bugs "unfixable" then claim they have to be fixed, which is a convenient oxymoron)

I think we can reason on the vast topics on private. I just gave you a hint: how do you explain the existence of C14 in diamonds that are supposed to be billions of years old?

First of all the measurement is probably within the margin of error, secondly you'd then have to discount external natural contamination, and lastly you'd have to discount internal contamination.

[sergiu]

Unfixable bugs are what is known in programming as design bugs for which there is a need for complete redesign which in turns adds different bugs. If you find out that design is flawed after having 100 million lines of code, that's a big ouch. Bugs cannot be escaped.
There is no margin of error for C14. It's found consistently in all diamonds tested. There are rumours of dinosaur bones which when cut, were found with soft tissue also, but that I would grant as rumour.

Side notes, about 10 years ago I heard rumours about an internal project, I think at IBM which was aiming at designing frameworks that could code full blown programs based on business analyst requirements... Kind of AI that could replace developers which are costly, complain and always migrate for better payments. 10 years later and I still have a job.

[MattiasNYC]

Unfixable bugs are what is known in programming as design bugs for which there is a need for complete redesign which in turns add different bugs. If you find out that design is flawed after having 100 million lines of code, that's a big ouch. Bugs cannot be escaped.

My argument wasn't that all bugs could be fixed. My argument was that if you think that bugs prevent an AI from reaching the level we're talking about because humans are incapable of fixing enough of the bugs (or are incapable of improving the design) that doesn't logically lead to an AI having the same limitations. It simply doesn't follow.

There is no margin of error for C14. It's found consistently in all diamonds tested.

I don't think you understood the explanation.

[sergiu]

@MattiasNYC
On paper we can gather the brightest minds and design an AI. Let's say we can prove that the design can work flawlessly as long as the implementation is perfect. That's fine. In practice, if complexity is high enough, there will always be a certain number of bugs. That's because with some exceptions, is almost impossible to prove an implementation and many times it's impossible to fully test it. Theory is perfect, I do not question that. Practice experience about implementation says something else. On top, for some approaches, validating the result is impossible.

Regarding C14, I would kindly ask you to inform yourself about the problem details. It's clear that you do not understand it or you do not have sufficient knowledge now to comment about it.

[MattiasNYC]

@MattiasNYC
On paper we can gather the brightest minds and design an AI. Let's say we can prove that the design can work flawlessly as long as the implementation is perfect. That's fine. In practice, if complexity is high enough, there will always be a certain number of bugs. That's because with some exceptions, is almost impossible to prove an implementation and many times it's impossible to fully test it. Theory is perfect, I do not question that. Practice experience about implementation says something else. On top, for some approaches, validating the result is impossible.

Ok. And?

Regarding C14, I would kindly ask you to inform yourself about the problem details. It's clear that you do not understand it or you do not have sufficient knowledge now to comment about it.

You should probably listen to actual scientists on that issue rather than some creationist sources. If you do you'll find that among serious, intelligent and accomplished scientists there is zero disagreement on the earth being around 4.5 Billion years old. Zero. Only uneducated misguided creationists have that belief.

The reason you trust technology every single time you get on an airplane, every time you get in a car into traffic, every time you file your taxes online, every time you use your ATM, every time you use your microwave.... the very same fundamental underlying reason is the reason we know the earth isn't "young". It's just ignorance to be a young earth creationist in the year 2017.

[Ket]

Given that skynet wanted to eradicate humans, I don't see how any reasonable person would want it in charge of humans

Don't let your extreme hate for a person blind you to everything, geez. I loath Hillary, she's the epitome of some of the worst political corruption, but I still wouldn't want Skynet over her.

Ouch.

My point is, a AI that learns for itself, puts tens of thousands of years worth of thinking in to each decision, and likely cross-references its own "thoughts" with the entirety of human civilisation (because it has no other reference), is clearly going to be making decisions for the long term benefit. One could conclude from this that the AI would either react as Skynet does, or see itself as a "caretaker", and take steps to ensure the planet, and humans, survive. In contrast humans pollute the planet, some countries with almost a complete disregard and ignorance, and make short sighted decisions that largely are made with ones own self interest in mind, of which half get reversed every few years. Humanity, at least in terms of those in power, are always 1 step forward 2 steps back. Nothing ever truly changes because theres always some idiot that comes along to piss it all up. I am not a believer in humanity, I have never seen enough proof to suggest it won't destroy itself through divided, small, insignificant opinion over important subject matter and denial / ignorance.

Oh and Skynet did not want to inherently destroy humans, it just concluded humans were a threat to it so acted to defend itself.

[sergiu]

@MattiasNYC
I think the best way to summarize my point is that the devil is in details. Just like it is theoretically possible to travel at speed of light, but for all practice we can only build machines that barely reach a fraction of it (and please do not come with Alcubierre drive).

I do see a an authoritative tone in the way my comments are picked up and attempts to ridicule them for not being politically correct. I do also notice superficial reasoning. For the purpose of education: radioactive dating is based on the following assumptions:
1) Closed system
2) Clock is set to 0
3) Decay rate is constant
Point 1. for all purpose is not true as water can wash away elements from rocks (fact). Point 2. is wrong as based on the conditions the material is formed, it can have more or less of parent or daughter isotopes (proven by observations). Point 3 can be observed to be constant in the last 100 years. That does not mean necessary it's constant. An analogy is to sit on a bridge, measure the flow of water for 1 minute and assume the flow of the river is constant over the complete year. The reasoning I used in the analogy is identical yet we both know we would fail a geology/geography exam if we would use such a reasoning.
Measuring the ratio of isotopes with state of the art mass spectroscopy is science. Making assumptions is not. Science does not forbid neither evolution, neither creation as there is no definite proof for any of them and will never be. It is man who forbids one or the other based on the world view he or she has. I'm entitled to interpret the evidence as best as I can and connect the dots (and of course be guilty for claiming one position): helium diffusion rate from rocks, even with 5 orders of magnitude of error is strong evidence which suggest radioactive dating is off and therefore game over for evolution. You see, planes are still working with a young earth, ATM are still functional, CPUs in 7nm are still possible, quantum computers are still working, surgeons can still do brain operations and we will still be able to send probes to other parts of our solar system. The only thing that changes is that the existential "is there a God" becomes "which God"?

I do apologize for abusing this thread with slightly offtopic and this would be my last comment on it, however we can reason over private messages.

[MattiasNYC]

First of all, you're still not understanding just why your argument about C14 being present is not invalidating dating using radiation. The contamination has the effect of carbon-dating NOT being used if the amount being measured falls within the error margin. So you're NOT arguing against the procedure.

Secondly, like other populist creationists you seem to be leaving out all the other radiometric methods of dating. Go up on Wikipedia for a quick look of the other 15-20+ methods used.

Thirdly, if you look at point #3 above you should see a problem.

You end that paragraph by saying that with a young earth view airplanes still work, but you fail to understand what you just said within a larger context. If there's no reason to assume that how the laws of physics act today was how they acted 1,000 years ago, then there's no real reason to assume they'll work the same a decade from now. Yet everyone contributing to the design of airplanes have managed to correctly foresee that the laws of physics did not change over the course decades.

- When the 747 was put into service in 1969, do you think that the designers figured "hey, let's just hope this plane will fly in 20 years, because who knows if gravity works the same then...."

- When they built the nuclear power plant that powers your home, do you suppose they sat down and figured "Gee, with this huge investment I sure hope that the atoms behave like they do today in a few decades..."

And no, I'm not ridiculing your comments for not being "politically correct", which is a silly comment to make. I'm just disputing them because the young-earth creationist argument is just nonsensical, irrational, illogical, and creating or spreading ignorance. And we have enough ignorance as it is on this planet and we don't need more of it.

[MattiasNYC]

My point is, a AI that learns for itself, puts tens of thousands of years worth of thinking in to each decision, and likely cross-references its own "thoughts" with the entirety of human civilisation (because it has no other reference), is clearly going to be making decisions for the long term benefit. One could conclude from this that the AI would either react as Skynet does, or see itself as a "caretaker", and take steps to ensure the planet, and humans, survive. In contrast humans pollute the planet, some countries with almost a complete disregard and ignorance, and make short sighted decisions that largely are made with ones own self interest in mind, of which half get reversed every few years. Humanity, at least in terms of those in power, are always 1 step forward 2 steps back. Nothing ever truly changes because theres always some idiot that comes along to piss it all up. I am not a believer in humanity, I have never seen enough proof to suggest it won't destroy itself through divided, small, insignificant opinion over important subject matter and denial / ignorance.

Oh and Skynet did not want to inherently destroy humans, it just concluded humans were a threat to it so acted to defend itself.

Yeah, I see your points.

I would definitely come down on the side of a future AI feeling the need to protect itself from us. We're not a reliable species in that way.

[sergiu]

@MattiasNYC
Is it ignorance to ask questions? Or is it ignorance to assume there is a naturalistic explanation for everything? You ignore the reasons and generalize! What on Earth has the possibility of having the decay rate accelerated in the past having anything to do with planes? Do planes run on Uranium? Is present impossible if there was accelerated decay in the past? Please check your private messages.

[Florinmocanu]

@MattiasNYC
The negation of a religious position without evidence that does not require the use of faith is still religion. No matter how you define it in dictionary.

Ehm, no .

So, as an atheist, i don't have to prove the "non-existence" of God because it's not an aspect of reality that has shown it's presence yet, only in human mind do you see it's presence, not in actual real and repeatable proofs/experiments. The human mind is not something you can trust, people can see/think/hear things under the influence of substances, under the influence of malnutrition/lack of water/starvation/deprived of sleep, under the influence of diseases. So, nothing that someone claims to have seen/heard can be held as an actual proof. A proof, a real proof is something which you can experiment upon, replicate and document, for others to be able to replicate themselves.
So, if you believe in GOD and demand proof from atheist, it's actually backwards. You should prove us atheists why us not believing in GOD is wrong, because you are the one believing in something which has no proof of existence, you are the one "holding the knowledge" over the existence of GOD.

So Sergiu, you can't request Mattias to prove to you that GOD is real. You have to prove to him such a thing. There is no such thing as proving a non-existence, it's logically fallacious.

God is real, because the existence of God cannot be disproved. That's your argument in a few words. Which is completely wrong/unscientific and has nothing to do with analyzing things logically. You see my point? If you want to show that GOD is real, you have to prove it, no us, atheists, disprove such a notion.

[sergiu]

A man once came home to his wife and said "Woman, I am a dead person, I am not alive!" . The man was sent to a team of doctors to try and convince him that he is alive. For months they showed him movies and experiments which had the conclusion that only people who are alive can bleed. Finally the man said "Ok, I know now as a fact that only people who are alive can bleed". In that second, the doctor took a knife and made a cut to the man's arm. And the man exclaimed: "Oh, dead people bleed too!"

The burden of proof trap is thrown every time when it comes to religion.
"God exists" - Can be true or false and by definition, cannot be proven. Why by default the community assumes "false" position?
"Radioactive decay rate was constant" Can be true or false and it cannot be proven for anything more than 100 years. Why by default the community assumes "true" as position ?

Does anyone sees something wrong in the statements from above? A position is taken in second case and based on inference, an assumption is made fact and the fact is used to support other assumptions. Just as second statement has observational evidence, the first one also has observational evidence, and let me illustrate with another example:

1. A 700MB data encoded in base 2 representing the source code of an operating system.
2. A 700MB data encoded in base 4 representing the DNA of a cell.

Both sets of data encode information and both are meaningless without an system designed to interpret and translate the data into useful actions. In the case of the operation system we have programming language, hardware architecture, input/output devices. In the case of the DNA we have the cell with its translators, replicators and other complex factories at nano scale. If we all recognize the first case needs a creator, why is wrong to assume that second case also requires a creator? The real answer is this: if we accept that information is created, it means we have to accept a creator and this goes against atheistic views. The real problem of atheism is accountability, not the denial of science. If there is a God, then you are accountable for your acts. Science by definition does not change if God exists. Fossils stay fossils, stars still stay in the sky. What changes is the basic set of assumptions used to interpret the evidence.
And before commenting that DNA is not information or that DNA can create itself, let me say this IS a religious position. And here is why: you choose an explanation contrary to all observational evidence. You can quote as many atheists, even a child of 5 years of age would recognize the problem in logic.

I'm fine with my belief in God and I'm fine with atheism. I'm here to reason, not to convert. What I am not fine with is the closed mind view of interpreting the evidence and suppressing any kind of questioning that does not fit the core assumptions. In this regard atheism in scientific community is the equivalent of fascism in political regimes. Richard Dawkins once stated the ultimate outcome of evolution is fascism.

[Ket]

Actually, Stephen Hawking has a equation that as good as proves there doesn't need to be a "god" for the universe to exist. If you have a deck of cards, infinitely shuffle it, just by pure chance you can draw all kings, or queens, or aces, etc. The point in that analogy is that you can, without any kind of hocus-pocus, come up with a "winning" combination, creating something from just base materials.

The more interesting question is why so much of the universe can apparently be explained in mathematical terms, it certainly does not mean there is some kind of mystical being, but if anything suggests there is, or was, a ancient civilisation that came long before we did, quite possibly even from a alternate reality, and created this universe with their advanced technology and understanding of things humans can't even possibly imagine or begin to understand. A sufficiently technologically advanced race would be able to do things that to anything less evolved would be incomprehensible and indistinguishable from magic, but it does not mean that race has any kind of mystical or divine "powers".

My thoughts on this question is that if humans were created outside of random chance, that advanced race is / was silicon based. Why silicon? Silicon reacts in extremely similar ways to Carbon based life, but here is where my theory really kicks in, if a silicon life form millennia ago created us and everything in our universe who is to say it was not already a artificial intelligence, capable of billions of years worth of problem solving in a extraordinarily short period of time? Something even beyond quantum computing. After all, what do we use silicon for right now? To create chips that are capable of insane amounts of processing power. Theres nothing to say that somewhere else in our universe, or a different universe entirely, that a silicon life form did not / has not evolved with these very same properties and over the course of their own evolution almost infinitely surpassed the abilities of what we know of as CPUs and GPUs.

The human brain is supposedly the most complex computer ever created, but what nobody thinks about is that the human brain is the most complex carbon based computer ever created, it doesn't mean there isn't a even more powerful, better, silicon based brain out there somewhere.