Comments by "justgivemethetruth" (@justgivemethetruth) on "Can we build AI without losing control over it? | Sam Harris" video.

  1. It's really doubtful that we can build AI, a general AI. Currently there is no such thing as self-awareness. The best chess playing computer in the world now or maybe forever has no idea what it is doing, what chess is, why it is playing it, and how to do anything other than play chess. It would not know how to talk about chess, or think about it. The AI thing, to me, is way overdone. The real threat to human survival is humans, because we do not seem to be intelligent enough to understand out own species or our own world. People reflecting their own dangerousness into a fictional machine, like projection, are ignoring each other and pretending AI is a threat.
    28
  2. Deep Blue is just more refined and complex version of regular programming, not the difference that would be needed to create actual intelligence. You just cannot let yourself get too wowed by all the fancy new technologies. It is like living in the age of horses and seeing a Ferrari and thinking if we can make something like a race car we will be able to make space ships that can get to the stars ... the two ideas are so far apart as to have almost nothing in common. We do not even know how to define or discuss intelligence. The only model of intelligence we have is anthropomorphic. All of man's inventions ... every single one of them have been copies of nature. We owe nature for everything we have ever done, and intelligence will be the same thing, whenever we can actually get a handle on it.
    3
  3. Well said, I completely agree. The energy we put into AI would be much better spent figuring how to get people to work together and stop warring. We don't have to worry about computers with mental capabilities or powers we do not have, but we do have to worry about people getting tools that can wreak havoc on our species and the planet - because we are anything but intelligence. We learned a lot of tricks by observing nature so far.
    3
  4. You are just talking about bugs in software, but intelligence, or general intelligence, involved goal setting, and learning, and re-defining goals, etc ... that is, we do not even know what it is, let alone how to define it or understand it let alone program it - and even less debug it. The complexities are so vast it is hard to imagine that we will ever be able to create anything but a clever encyclopedia. We can program it to make war for example, and maybe accidentally turn it against us, but it has no reason to live, it has no friends, it cannot conceive of any of this stuff, which we inherited from all the living life forms that evolved ahead of us, things built into our DNA from maybe as far back as millions of years that are so subtle that we cannot even see evidence of. The idea of AI being a threat is so silly compared to the idea that we are the threat and we have to figure out our own intelligence, or how to grow into it, because we are behaving like mindless robots ourselves.
    2
  5. -- What kind of an argument is that? What kind of a question is that. You stated that a computer could "read" a book ... so I think it is up to you to define what you mean by read? In what way can a computer understand or enjoy a book. Part of reading for humans, the only intelligent species we know is to understand and enjoy a book. The human brain is also not programmable in the sense of a computer. In a computer programming means the computer reacts the same way to certain inputs ... the human brain does not work like that. Define Intelligence?
    2
  6. Dude, write a book! ;-)
    2
  7. -- And as long as the conditions are in place, which they are, self conscious AI is bound to happen. Hey Froggie, you might believe that, but that is faith, not science, not reason, and you have no reason of rationale to think that, it's just hopeful wishing. It's like saying that we will figure out how to do faster than light travel. It is something different than any other kind of motion, and also has the addition problem of relativity effects so that even if we figure out how to do it, it may not be useful to us. When you study the problem with a skeptical mind and think about it there are a lot of problems to be solved just to define the problem and no reason to think we can do anything but perhaps single out small systems in the brain, such as the eye to copy and use in some design, not to determine how intelligence really works. Also, there is even less reason to think that if we can reproduce of create human consciousness that it would progress any faster than people do. We tend to think that because computers are so much faster at computing than we are that they will be able to make inventions and discoveries faster too, but any intelligence will have limitations and discoveries are made by how fast things happen outside, the right discoveries or inventions, not how fast people think.
    1
  8. -- they are trained to perform a task. And they can be trained to to almost anything: In this context, "training" is simply another word for a different type of programming. It is not training in the human sense. And to read 10 books a second is not reading in any human sense of the world.
    1
  9. You clearly have lots of experience with that ... and trolling.
    1
  10. I have simply said for an extraordinary claim extraordinary proof, not handwaving and bullshit is necessary. Now run along and pester someone else. The complexity of flight is nothing compared to the most complex arrangement of atoms in the universe. You know nothing, but you keep talking ... I wonder why?
    1
  11. The place where we are now is not equivalent to where we were a century ago .... the technology and inventions we developed in the last 100 years were obvious direct rip-offs from nature mostly, but the complexity of the human brain is in a completely different level. We do not even have the words to define intelligence.
    1
  12. I like the reformat our thinking phrase ... the thing is that as we reformat our thinking to be more precise and clear, we change our behavior and motivations, and without a social structure we are all fighting ourselves in ways that are more perceptible and analyzable - if we have education. Meaning that if intelligence does not have some kind of "programming" it it to get along, respect, cooperate or tolerate other intelligences - all it will get it war and destruction ... and it seems to be that by definition - that is not intelligence. Or another thing is that one intelligence that dominates everything is also not intellgence, even though that might be what logic tells a cold dispassionate intelligence to do. It is like after WWII, the finest minds got together and tried to reason as to what would be the best course of action for the US to take since at the time we were the only ones with nuclear weapons. The logical choice was to use the nuclear advantage to take over the rest of the world - but no one did that. No one also created rules for all countries and people to get along either though. To get to intelligence, something has to experiences or understand all of the other paths, and have a certain attitude about which path to take ... dominance, destruction or cooperation. We as a species really do not understand this. How do you have a main loop of purpose for a an intelligent being that is not motivated by some goal, and then even at Terahtz speed, how do you explore all of the paths of behavioral choices, because you have to do something? There is understanding, finding out, and then there is existing and wisdom ... maybe what we would call religion. There is some kind of purpose to religion that is analogous to an enclosing programming loop where a meta-understanding or even some kind of faith is necessary - I think,
    1
  13. The only example in the universe of intelligence, general intelligence, that we have is the human brain, so I think you ought to reconsider your comment about being self-aware not being important.
    1
  14. No, odds are it needs to be self-aware to be truly intelligent. But I am so glad you seem to know it all, maybe you should get a job with DARPA instead of trying (unsuccessfully) to convince me here online.
    1
  15. You fail to listen, you fail to think, you fail to understand. It really doesn't matter because it is not going to happen. You are really not worth talking to.
    1