Sam Harris' Ted Talk On The Future Of Ai

[Siim Land]

What do you guys think about Sam Harris' new TED talk?

Basically, if we accept the idea that intelligence is a matter of information processing, then it's inevitable that we'll eventually build artificial technology that will exceed us in everything, including ethics, moral etc. In the end, we'll be simply overcome and doomed as a species. This is quite probable.

What he wants us to think about is how we could build AI safely so that we wouldn't be destroyed by our own creation. It's definitely one of the most important questions of our generation. We're literally moving towards either our impending doom or salvation. 

Thoughts?

 

[ChimpBrain]

I think he and Bostrom are spot on with this line of thinking . There was another recent thread on a similar topic that got major pushback from members of this forum though, including Leo.  Frankly it seems completely inevitable to me assuming we don't blow ourselves back to the stone age within the next 100yrs. Who am I though ? Just a dude who read some books and watched some videos. 

 

OP if you like this topic check out Nick Bostrom's book Superintelligence . 

[Siim Land]

@John Flores Wow, that's just baffling and scary at the same time. We can create artificial consciousness. Gives some things to think about concerning the future of our own. I wonder whether the answer has any objective ground to it  But then again, it answered "what is 10 minus 2?" - 72.

@ChimpBrain Yes, Bostrom is amazing, I just picked up his book. Synchronicity

It's funny though. AI clearly has more intellectual potential and the development of our species is very much dependent of our technology. Our biological bodies are like outdated hardware, whereas the software, our consciousness, is willing to go through a quantum leap. 

Edited October 17, 2016 by Siim Land

[ChimpBrain]

55 minutes ago, Siim Land said:

 

@ChimpBrain Yes, Bostrom is amazing, I just picked up his book. Synchronicity

It's funny though. AI clearly has more intellectual potential and the development of our species is very much dependent of our technology. Our biological bodies are like outdated hardware, whereas the software, our consciousness, is willing to go through a quantum leap. 

I actually didn't know he had a second book, thanks! Unfortunately it doesn't appear to be available in audio format yet.

[Siim Land]

23 hours ago, ChimpBrain said:

I actually didn't know he had a second book, thanks! Unfortunately it doesn't appear to be available in audio format yet.

@ChimpBrain He doesn't have a book called synchronicity, if that's what you think. Synchronicity is just a meaningful coincidence. 

[ChimpBrain]

2 hours ago, Siim Land said:

@ChimpBrain He doesn't have a book called synchronicity, if that's what you think. Synchronicity is just a meaningful coincidence. 

Haha my bad man I misinterpreted what you meant by mistaking a period for a comma! 

[ChimpBrain]

@Siim Land This is another good discussion of AI by Sam if you're bored.

 

[Siim Land]

On 16/10/2016 at 0:00 AM, ChimpBrain said:

I think he and Bostrom are spot on with this line of thinking . There was another recent thread on a similar topic that got major pushback from members of this forum though, including Leo.  Frankly it seems completely inevitable to me assuming we don't blow ourselves back to the stone age within the next 100yrs. Who am I though ? Just a dude who read some books and watched some videos. 

 

OP if you like this topic check out Nick Bostrom's book Superintelligence . 

@ChimpBrain Just listening to Superintelligence and the shit is looking even scarier Amazing book and Bostrom's arguments are probable in almost any way you look at them. I love the part that this super AI will hide its plans of becoming a god from us until its too late. Also, the part where it could get access to protein replicators, have it shipped to its location by a FedEX guy who has no idea about it and then build itself an organic body. WHAT?! Maybe I understood it wrongly but it seems possible with recursive intelligence and cognitive abilities.

[Siim Land]

i think we're making the mistake of thinking that we know how this recursive AI is going to think and behave. Humans are cybernetic organisms, which means that we can adjust our own behavior. However, we're not completely so because we're still mostly governed by our emotions, thoughts etc. On the other hand, those things aren't guaranteed in a purely bayesian superintelligence. We can only predict what the actual situation is going to be because our mode of being is quite limited. Also, we can't underestimate the AI-s ability to influence human motivation and direct other people to do certain experiments for it.