Why A.I is impossible

Started by
116 comments, last by Alexandra Grayson 6 years, 1 month ago
9 minutes ago, Oberon_Command said:

13 trillion / 323 million is a bit over 40k. I'm not aware of any UBI concepts where participants are to be paid $40k per year. Pilot programs are typically paying something in the $16-24k range (eg, the one currently underway in Ontario) and I would assume that a full-scale program would be similar. I don't know where you got that $40k number, but I suggest that your dismissal of the concept is founded in this assumption and not any realistic conception of how UBI would actually be implemented.

Well, it would have to be implemented in a way that gives a decent quality of life in a situation where there's massive unemployment, right? Add in the fact that health insurance in the USA is prohibitively expensive when not employed (My wife has an ER bill for $30,000+ I've been refusing to pay for years now), Adjust it for inflation for a decade, and assume it's for the "typical" household of 2 adults 2 children (replacement rate, basically), and $40k~ is way closer to it. If there was mass unemployment and only $24k~ guaranteed income, there would be mass riots/revolution.

 

This is also speculative to the future, population will be higher if/when it's actually implemented, so there's even greater costs than with the 323 million people estimate. Currently you could knock off around 40m~ people from that estimate for underage/illegal immigrants.

Advertisement
23 minutes ago, conquestor3 said:

Well, it would have to be implemented in a way that gives a decent quality of life in a situation where there's massive unemployment, right? Add in the fact that health insurance in the USA is prohibitively expensive when not employed (My wife has an ER bill for $30,000+ I've been refusing to pay for years now), Adjust it for inflation for a decade, and assume it's for the "typical" household of 2 adults 2 children (replacement rate, basically), and $40k~ is way closer to it. If there was mass unemployment and only $24k~ guaranteed income, there would be mass riots/revolution.

It seems reasonable to assume that in a world where the US actually implements UBI, it has also implemented universal healthcare, just like every other developed nation, so the costs would be significantly lower. I don't think anyone would disagree that universal healthcare is politically more likely than UBI.

(Seriously, $30k for an ER visit if you're uninsured? In Canada, even if for some reason you don't have provincial health insurance, an ER visit would be at most $600. What the hell is your country smoking?)

Anyway, this thread isn't just about the US and it's myopic to pretend that assumptions that apply to the US as it is right now invalidate a concept that would be applied world-wide in the future. I also doubt that UBI would actually be implemented at above poverty level numbers regardless of the quality of life that provides.

20 hours ago, conquestor3 said:

The 2 are related though, an intelligent AI would be able to be applied to solve automation pretty quickly. A smart enough AI would be able to immediately automate intellectual jobs as well.

I don’t disagree at all, and automation is a great topic worthy of discussion.  

But this thread is about the possibility of AI, not the potential effects. Besides, automation will happen with or without AGI. I feel like it would be better in a separate thread. 

if you think programming is like sex, you probably haven't done much of either.-------------- - capn_midnight

Setting aside the whole spirituality/soul angle, "human-like" AI exists because humans exist; we are walking examples of ourselves.

If the human body is sufficient to create a human being, then we can clearly create AI.  That's not to say it's a guarantee that souls don't exist or can be emulated, but to assert that would take us outside the bounds of provability and into belief.

 

Mind you, there's a far cry from creating people and creating the whole "self-improving superhumans" that most quacks "futurists" worry about.  In that regard I wouldn't call it "impossible", but I certainly see no reason to believe such a thing is possible.

19 hours ago, SeraphLance said:

"human-like" AI exists because humans exist; we are walking examples of ourselves.

That makes no sense at all. Unless you're going to redefine AI to include organic intelligence, at which point this conversation becomes pointless. 

 

19 hours ago, SeraphLance said:

If the human body is sufficient to create a human being, then we can clearly create AI. 

Except clearly we can't, because we haven't (not yet, anyway).

 

Creating humans is easy (in fact, it seems difficult to stop creating humans). We're talking about machine intelligence that doesn't have the baggage of human intelligence (i.e. perfect memory, perfect attention span, perfect accuracy, etc).

Now, it would be interesting if it turned out that our imperfect intelligences are inherent emergent properties of creating a human level intelligence, and that any machine intelligence we build will ultimately end up replicating the flaws.

if you think programming is like sex, you probably haven't done much of either.-------------- - capn_midnight
4 minutes ago, ChaosEngine said:

That makes no sense at all. Unless you're going to redefine AI to include organic intelligence, at which point this conversation becomes pointless.

If I created a human being from whole cloth in a lab, I would have created an artificial intelligence.  Whether AI is "organic" or "mechanical" has no bearing over whether or not it is AI.  Aren't you an Atheist?  I would think that outside of spiritual beliefs, everyone ought to think creating an AI is possible this way, even if we can't realistically do it now.

I don't think it's a terribly interesting observation, honestly.  It's kind of axiomatic, even.

1 hour ago, SeraphLance said:

If I created a human being from whole cloth in a lab, I would have created an artificial intelligence.

1: we can't do that.

2: not really. You're just spawning another instance of an already existing class of intelligence. The mechanism you used to create it might be artificial, but the end result is "natural". 

1 hour ago, SeraphLance said:

Whether AI is "organic" or "mechanical" has no bearing over whether or not it is AI

Wikipedia thinks it does. "Artificial intelligence (AI, also machine intelligence, MI) is intelligence displayed by machines, in contrast with the natural intelligence (NI) displayed by humans and other animals. "

So does Webster and Britannica

I agree that we might end up creating an intelligence by organic means, but saying that creating a human is creating an AI renders the discussion pointless. 

1 hour ago, SeraphLance said:

Aren't you an Atheist?  I would think that outside of spiritual beliefs, everyone ought to think creating an AI is possible this way, even if we can't realistically do it now.

Being an atheist has nothing to do with this. For the record, I do believe it's possible to create AI, but I'm also willing to accept the possibility that human-level intelligence is some unique, as-yet-undiscovered inherent property of organic intelligence.

if you think programming is like sex, you probably haven't done much of either.-------------- - capn_midnight

The argument in this thread assumes some form of vitalism. I don't understand the appeal of this way of thinking. All the evidence we have is that there is nothing special about living organisms: They follow the same laws of Physics as everything else.

If we reject vitalism --as we should--, the existence of human beings is proof that intelligence can be achieved by physical means, so there is no a priori reason why a machine couldn't be intelligent.

EDIT: I found a Wikipedia page about the subject: https://en.wikipedia.org/wiki/Hard_problem_of_consciousness. The last paragraph quotes Stanislas Dehaene drawing a parallel to vitalism. I wholeheartedly agree with him.

1 hour ago, alvaro said:

If we reject vitalism --as we should--, the existence of human beings is proof that intelligence can be achieved by physical means, so there is no a priori reason why a machine couldn't be intelligent.

I tend to agree with this. Whether humans are smart enough to build one is another question.

1 hour ago, alvaro said:

EDIT: I found a Wikipedia page about the subject: https://en.wikipedia.org/wiki/Hard_problem_of_consciousness. The last paragraph quotes Stanislas Dehaene drawing a parallel to vitalism. I wholeheartedly agree with him.

A machine being intelligent and a machine being conscious are two different things. A machine could theoretically be intelligent without having a subjective experience or a sense of self. It could be that consciousness is a unique emergent property of organic intelligence (I personally don't think this is the case, but it's possible).

The first chapter of Greg Egan's book Diaspora describes the formation of a new sentient AI entity. It's fiction, but it's got some interesting ideas and is worth a read.

if you think programming is like sex, you probably haven't done much of either.-------------- - capn_midnight

"Consciousness" and "Intelligence" are big words which could mean different things in different contexts.

As in any philosophical debate, it is worth spending time defining the terms being used in the conversation. I would argue that defining consciousness constructively is a dead end. So instead of defining "all things conscious",  or "some things which might be conscious": Let us take the destructive approach and try to define: "Some things which are intelligent but not conscious".

Let's assume that living organisms can be vaguely sorted by intelligence. For simplicity's sake let's forget the special cases of dolphins and octopuses.

Let's rate them from "stupidest" to "smartest"

1. Single cell (Germ)

2. Small Multi-Cell 

3. Nervous system (Jellyfish)

4. Worms

5. Fish

6. Reptiles 

7. Mamals

8, Social mamals (dogs?)

9. Humans

At what point would you say it is intellegent but not concious for certain? At what point would you consider it a "maybe"?

I would be interested to hear the reasoning behind your choices.

 

My Oculus Rift Game: RaiderV

My Android VR games: Time-Rider& Dozer Driver

My browser game: Vitrage - A game of stained glass

My android games : Enemies of the Crown & Killer Bees

This topic is closed to new replies.

Advertisement