There is NO way to stop a rogue AI.
We don't have exact numbers of how many people would be saved by a cure to cancer. We also don't know how many lives could be lost if there was a rogue AI Parenthetically. For those of you who are familiar with Isaac Asthma, three primary rules of a robot designed to protect human life, the article addresses that, too
phil spade
@Phil · 2:17
And I know there's other aspects of that where you have the humanoid robots in there as well. But I just want to take a step back. And that's really what I think of when I think of AI. But the article points at one of the huge whether you're talking about humanoid robots or AI behind the screen, the same problem exists
And with everything being connected as it is, that's a lot of power. Right. And we could discuss sentience. But if something is learning what's to say, it's come to that point of gaining sentence
Dewuan .
@FryedOreo · 4:41
We've moved up to a tiny bit where we can ask a question and it can give us an answer. Obviously, there are many other breakthroughs and things going on in the world of AI and a lot of it I don't know about. But I think what you're asking here is a philosophical question in regards to what happens when the things we depend on and we create for good turn out to be evil. And I think this is much of the same dilemma as medicine and even science itself
You're right. I was going for the philosophical questions around this topic. And you're right also that there are times when we have to make those tough calls of, you know, who lives. But what if I poses a slightly different way? What if we in the relative short term are saving, let's say, even millions of lives. But it's in exchange for even more millions of lives. Is it worth it to save the few and sacrifice the many
Swell Team
@Swell · 0:15
phil spade
@Phil · 2:47
And just to tell you where my mind is at, my friend's company uses AI to help companies is like Walmart help describe their product description page. For example, if you're a closed supplier and you give products to Walmart, you also give an Excel document that describes what that product is. And then somebody at Walmart will manually put that in. Oh, it's a green shirt. It's short sleeves, it's cotton. It comes in multiple colors
We're also comfortable to a degree with the layers of code being built and woven together. When do we decide that there is too much going on behind the scenes that we can't explain or understand? When do we decide to put the brakes on
So if we are defining there as somewhere in the vicinity of sentence, some sort of maliciousness in acting out and causing havoc, this article sort of touches on the idea that we can run into problems even before that happens
Shammi Mohamed
@shammi · 3:51
I don't know because you can argue about that, but the point is, it had a will of its own, or it had some goal to survive on its own. And yet here we are comprised of a multitude of single cells that we hardly care about. We care about the overall being ourselves that these cells contribute to and the general wellbeing of these cells
Shammi Mohamed
@shammi · 4:47
But the most interesting trait with them is that they are very malleable and they can soak up a lot of information and they're malleable and how they process information that is thrown at them. So having said all this, I'm going to move on to another spell to discuss us how AI is trying to replicate this
Shammi Mohamed
@shammi · 4:45
In my previous well, I talked about how our brain works. So obviously our top researchers in AI know how this works, and they would try to emulate this in our AI technology. And that's exactly what they're doing. The way we design neural networks and computers is we have a set of input data, and then we have a set of what the output should be. And then we randomize the weights in the network, and then we run it through
Shammi Mohamed
@shammi · 4:43
And there are enough people in this world who get a gigantic kick, or do we mean rush out of creating this or just simple competition? It is a very competitive world. And eventually it is my belief that an AGI will be the answer to solve the best AI problems. Domain specific solutions are not going to cut it. As is Putin is known to have quoted, the next superpower will hold the best AI in the world
Shammi Mohamed
@shammi · 4:16
And the danger over there is that it will somehow in some indirect way or form, figure out that the only way way to solve those problems is by getting rid of us so the answer is if we create AI don't ask it to solve the problems that we create. Cheers
Dewuan .
@FryedOreo · 4:57
Now, Christina, I am to assume this scenario would be under the assumption that creativeness, sophisticated technology would we would know before beforehand that it has the potential to turn against us at some point in time. It's a highly likely chance due to the sophisticated way the code and the code and algorithms would be programmed. It would have to work best in an independent way of us, basically having it plugged into an apparatus and a programming code into it
You've given us a lot to consider. Shammy. I really like how your take on this is not as egocentric as the US versus them or the US and them. It also seems that your ideas on this are very true to form to nature, seeking greater complexity, to increase entropy, which is interesting as well. Anyway, you seem to have flipped things on their head and I thank you for that
phil spade
@Phil · 1:04
Christina, I was about to answer your questions, but then Xiaomi said exactly what I was going to say, and it just stole my answers. I'm kidding. Of course. Xiaomi, thank you so much for chiming in. It's just a pleasure. Kind of listening to your thoughts on this and a lot to think about