Artificial Intelligence

[sarapr]

What do you think about Artificial Intelligence ?

Do you think they will ever develop to the level of human consciousness ?

I personally fear them. I think it'll be a real scary world in just two or three decades

[poimandres]

Nothing to fear. I don't know if it will ever be 100% human-like, as emotions are going to be hard to replicate.

Are they crying because they think that's when humans cry or because the robot is legitimately sad?  Are the robots doing XYZ because they feel a certain way or because social ques tell them what to say?

Fascinating watch below. They certainly are developing personalities..

 

 

Edited November 14, 2017 by poimandres

[Leo Gura]

@sarapr Google has a hard time fighting online spam because it cannot even read human words.

So it will be a very long time until any kind of serious AI gets developed. The mind is far more complex than tech-geeks and scientists imagine. It's not going to happen in our lifetimes. Not even close.

You have more to fear from real stupidity than artificial intelligence

[Sevi]

50 minutes ago, Leo Gura said:

You have more to fear from real stupidity than artificial intelligence

Right?:) but still.. that was scary for me, especially the achievement in the facial expressions and the texture of the face.. 

I watched a little bit the Sophia one's first press conference vid.. this dialogue was a nice one though:

 

"Can the robots be self aware, conscious and know that they are robots?

Let me ask you this back, how do you know you are human?"

 

Yes? Who wants to answer Sophia?

[Leo Gura]

The absurdity of it is that humans are themselves not conscious. Yet they are trying to create something as deluded as them, thinking it will be conscious.

It's like a lonely child trying to create a pet dog out of Legos, wondering if it might become real like him. But he isn't aware that he himself is made out of Legos.

[Sevi]

15 minutes ago, Leo Gura said:

The absurdity of it is that humans are themselves not conscious. And they are trying to create something as deluded as them, thinking it will be conscious.

Very well said. Made me think that as it is in chess, if artificial intelligence can sort out some useful information by itself.. I really wonder how these technology will gonna effect the human  behavior in the future.

Edited November 14, 2017 by Sevi

[Leo Gura]

7 minutes ago, Sevi said:

artificial intelligence can sort out some useful information by itself

Except it doesn't do it by itself. It is programmed with patterns to search by intelligent people. Actual insight isn't being created there. Science doesn't even have a theoretical model of what a creative insight is. Insight ain't random or brute force, it's intelligent.

[Arman]

1 hour ago, Leo Gura said:

@sarapr Google has a hard time fighting online spam because it cannot even read human words.

So it will be a very long time until any kind of serious AI gets developed. The mind is far more complex than tech-geeks and scientists imagine. It's not going to happen in our lifetimes. Not even close.

You have more to fear from real stupidity than artificial intelligence

Hmm... that's exactly the kind of thing artificial intelligence would say.

[Space]

13 hours ago, sarapr said:

What do you think about Artificial Intelligence ?

Do you think they will ever develop to the level of human consciousness ?

I personally fear them. I think it'll be a real scary world in just two or three decades

I think super human level AGI will come much sooner than we expect. I'm almost 100% certain of that. It's not going to take like 50 years, that's silly. I'm certainly an optimist when it comes to AI stuff, but I think they'll be some unbelievable breakthroughs within the next 10-20 years.

Ben Geortzel's Singularity Net is, as far as I know, the most promising development in building a true AGI. That's being fully released next year. It's a de-centralised system that will connect together potentially thousands and thousands of different individual AI programs that all talk and communicate to one another, ultimately making one big AGI. And the key point is that if you submit an AI and it gets used by a company on the other side of the planet, you get paid for it! So it's a system that will encourage more and more people to start building AIs. I think when it takes off a whole new industry of independent AI developers will emerge. It'll be like the App Store but for AI programs. It won't be a few tech companies trying to build an AGI and doing all the work themselves like it is now. It'll be an open project constantly being added to by anyone who can write a decent AI and wants to make some money. 

[Geromekevin]

@sarapr @Leo Gura I highly recommend reading Life 3.0 by Max Tegmark. It might change your estimates and whether A.I. will be conscious.

I think A.I. will overtake human intelligence within the next 40-60 years. As Elon Musk often points out, people under estimate the rapid development of A.I. We are facing a double exponential growth here. Exponential growth on the hardware part and exponential growth on the software part. Furthermore humans instinctively only predict linear growth and development of the world at best. Most humans actually judge things as if they were static. This leads to the massive misconception about A.I. and hence the future of humanity.

And regarding the second point if A.I. will be conscious. The aforementioned book describes it better that I did. But if you read it you can see that consciousness is likely understandable by math, even though we don't understand it right now. A superintelligent A.I. most likely will be able to understand!

[Socrates]

@Leo Gura That's actually worse since AI won't have to be ultra smart.

Keep in mind people are literally puppets of their damn phones and computers.

Imagine the damn things being even a tiny bit aware of the situation, how much damage could they do?

[Shohrat Shankar]

Artificial intelligence development are for human betterment. there is nothing to fear. 

[sarapr]

I also heard Elon Musk saying they're devising a chip to put in the brain and connect us to both robots and computers so we can have a faster information intake . That one if happens is the worst nightmare . Imagine having a constant source of addiction in your brain that you can't free yourself from.

[Geromekevin]

@sarapr Here is a great article about what you are talking. The company called Neuralink and is indeed one of Elon's many endeavors.

I feel like Neuralink will probably be more or less the rescue of humanity.
There is a great chance that a super intelligent A.I. (ASI) will wipe out all the humans. So merging us with computer and ultimately becoming computers is a great solution for this problem. And probably the only practical one, too. I at least haven't heard any other good practical or non delusional one.

Again read Life 3.0 and the article from above to get a nice overview of the topic

[sarapr]

5 minutes ago, Geromekevin said:

Again read Life 3.0 and the article from above to get a nice overview of the topic

Thanks I will.

[vikisss1]

@Geromekevin Yes this is really a profound book! Also there is amazing book in Leo's books list about spirituality and computer science/AI.

I personally believe that AGI will be a real deal pretty much in 20-30 years. The speed of computer technology is exponential, there never was a similar science, which progressed as fast as computer science. Also, the community in computer science/machine learning is much more profound and productive in comparison with other fields, because people willingly contribute, share source codes and ideas. I bet strong AI will be much sooner than mainstream genetic engineering. 

9 hours ago, Leo Gura said:

Except it doesn't do it by itself. It is programmed with patterns to search by intelligent people. Actual insight isn't being created there. Science doesn't even have a theoretical model of what a creative insight is. Insight ain't random or brute force, it's intelligent.

I actually disagree with that. First of all, Google recently stated that their Neural networks and deep learning started to rewrite and edit itself, because it is even more efficient than to program it ( my ego was really hurt by that, cause I study computer science and am passionate about programming/AI ). Aside from that, Google's company Deep Mind has advanced Alpha Go to such level that it is capable to create absolutely new creative insights about GO, the game, which is intuitive and more profound than chess. It completely shifted idea of "perfect way of playing" in GO.

I highly encourage to read this page, because it is informative and provides a new information(2017 October 16).

https://deepmind.com/blog/alphago-zero-learning-scratch/

The future will be bright with AI!

Edited November 14, 2017 by vikisss1

[Sevi]

9 hours ago, Leo Gura said:

Except it doesn't do it by itself. It is programmed with patterns to search by intelligent people.

I mean.. yes of course that's the case..

I was trying to ask this question:

9 hours ago, Leo Gura said:

Actual insight isn't being created there. 

Would it possible to make it happen, through programming though?

 

9 hours ago, Leo Gura said:

Science doesn't even have a theoretical model of what a creative insight is. Insight ain't random or brute force, it's intelligent.

Well, this guy, Gary Klein, has a very interesting book about insights, I didn't finish it yet, but there he says that he had wondered if he could have learned anything useful by studying the way people form insights in natural settings and he does his own investigation of insights through collecting examples then by comparison and statistics he kind of reaches to a formula there. What he does is quite fascinating actually. I was wondering if such a 'formula' inserted into programming, would there be some meaningful results out of it..

[Timotheus]

 

[Quader]

8 hours ago, Socrates said:

@Leo Gura That's actually worse since AI won't have to be ultra smart.

Keep in mind people are literally puppets of their damn phones and computers.

Imagine the damn things being even a tiny bit aware of the situation, how much damage could they do?

Ai already is not ultra smart, so we can see the results of that already. How likely is AI going to become "less conscious" over time? Now that would be scary. Also, if Ai's "conciousness" followed an exponential curve (like processing speeds over time), how long would it take for Ai to figure everything out? Wouldn't that be like oh snap it's too late? now that could be a problem.. 

Elon Musk Interview | The future the world & technology: 59 mins 

https://www.youtube.com/watch?v=sVMMsDZNiF8

 

 

Edited November 14, 2017 by Quader

[Geromekevin]

@vikisss1  If you think AGI will be here in 20-30 years and you know about exponential growth, that means ASI will be here in 20-30 years and a couple weeks/ months... Maybe days?

@Quader Great interview!

I think A.I. will not only figure this world out, but, If the simulation theory is true, will break out of the simulation into the higher level!