1. Why don't you Sign Up? It's free and doesn't require anything special.
    We are a friendly community and you might be much smarter than you think.
    Don't just look at us. We need YOUR help. Let's build together the next big High IQ Society.
Dismiss Notice
This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

The unfathomable machine

Discussion in 'Computers and Technology' started by Tiaric, Jun 10, 2017.

  1. Tiaric

    Tiaric Active Member Claimed IQ: 130+

    Do you think we will be able to build a computer one day, that we won't understand the inner workings of, although we will understand the output, the machine will have learned to communicate with us?

    Or are we close to doing that now?
     
    Cornucopia and mfrittman like this.

  2. TheGreat

    TheGreat Member Claimed IQ: 130+

    I really don't understand what you mean. First of all, you can't build something you don't know how it works unless its components are not man made but, Alien made, which is not possible because aliens are not scienfically in existence. Creativity comes from understanding then, innovating. You can't skip the understanding part. For example, we needed to understand why a bird could fly in order to build an airplane.
     
  3. bitcoinmonster

    bitcoinmonster Active Member Claimed IQ: 130+

    That is not true. Developments accelerate, and so at some point or another, one person will not be able to understand all the inner workings of a given system.

    And I have one example already: the brain. Yes, for all intents and purposes, we built it because we are it. We understand the output but not the inner workings.
     
    Cornucopia likes this.
  4. Tiaric

    Tiaric Active Member Claimed IQ: 130+

    Well, we already build "machines" that can learn, right? We don't let them do this in isolation, which is to say, we maintain communication with the machine or we essentially teach it how to tell us things.

    Suppose we can build a machine one day that can learn in a way we don't understand, although we understand how the machine is put together we no longer understand how it works. Not just some people, but everyone. Eventually we forget the details of the machine's design and machines do all the designing (of their successors) for us.
     
    Creedinger and TheGreat like this.
  5. TheGreat

    TheGreat Member Claimed IQ: 130+

    first of all, the brain is not a machine, its an organ. Just because our brain reflects our personality doesn't necessary mean we built it. In fact, no body organ can be classified as man-made or that we have voluntarily control over. In addition, we understand most of the inner workings of the brain and we fully understand the output. don't forget the auditory nerve, olfactory nerve, optic nerve, taste buds; these are nerves that sends information to our brain(so, you see we know most of the input) our brain then processes it, using our cerebrum,(thinking part of our brain) medulla oblongata(involuntarily control) or the cerebellum( for motor movements) considering which one is needed at certain time. and then sends the output, through nerves to various part of the body. so, you see. we understand mostly how the brain works on a fundamental level. (maybe not on an advanced level) like how our brain has to do with our sexual orientation(which is still an ongoing research)
     
  6. TheGreat

    TheGreat Member Claimed IQ: 130+

    Thumbsup! I have no choice than to agree with you because this is very possible with robots. we can create them and give them a very large storage capacity.this way, they are likely to learn and think outside of the box. I get what you are saying! :)
     
  7. mfrittman

    mfrittman Active Member Claimed IQ: 150+

    Perhaps future manufacturing and design will be handled by AI. After a few generations, we will likely build machines we can't understand.

    My wife built a brain in her womb.
     
    Moloch likes this.
  8. Moloch

    Moloch Well-Known Member IQ: Over 150

    Organs are constructed of cells which are constructed of proteins which are constructed by DNA, in essence, which is a computer code consisting of 0, 1, 2 and 3. Organs are a hive of cells working together, all seperate units. Brains are self-altering neural networks. The difference between a machine and an organ is just the building blocks. Is a computer a machine? Is a bio-computer a machine?

    The joke is thus: Mankind seeks to replicate itself in the form of an electronic machine as it considers this to be a superior form, whilst mankind already has a self-replicating, intelligent pattern-recognizing machine.. itself. If we were to harnass the true power of the human race then we would be able to solve many things in no-time.
     
  9. bitcoinmonster

    bitcoinmonster Active Member Claimed IQ: 130+

    An organ is a machine because a machine is simply "an apparatus using or applying mechanical power and having several parts, each with a definite function and together performing a particular task" (from Google!). And don't kid yourself-- we know next to nothing about the definite nature of the brain.
     
  10. TheGreat

    TheGreat Member Claimed IQ: 130+

    Alright then, what do you think is the definite nature for the brain?
     
  11. Creedinger

    Creedinger Well-Known Member IQ: 120+

    I think the idea at the moment is making human workforce expendable as it is quiet expensive.

    In our current economic system which drives the demand for innovation, the value of a person is estimated by the presumed output in relation to cost for the investor and here humans just suck compared to computers and robots.

    In a few years there wont be a back office anymore as all accountants, controller, Human Resources, administration will be automated. In case the ai is open source this is no problem but the ais currently most advanced use big data for training so google facebook et al. Are the only companies, which can utilize this as they own the data.
     
    Last edited: Jun 11, 2017
    Moloch likes this.
  12. conrad

    conrad Member Unclaimed IQ

    Some poor computer programmers have to endure that situation already. Either a colleague of his wrote a program which code is,by any stretch of the imagination and logic, absolutely and simply unfathomable, an intricate mesh of unlogical tangle, an impenetrable jumble of GOTO's, loop and variable names without any semantic meaning whatsoever - or there is a bug deep within the system, a complex systems of cogwheels the inner and mutual workings of which noboby in the company understands. because it has developed its own dynamic, a bug occuring only in extreme cases which aren't reproducable. Then, you get a hunch of the power and the terror of automomous AI first hand. I've been through that hell two times.
     
    Creedinger likes this.
  13. Creedinger

    Creedinger Well-Known Member IQ: 120+

    So you have worked on a team project twice already? Btw you can usually delete all the green parts of the code and it still works lol.

    I love it when programmmer 1 writes the code and then programmer 2 adds the comments. A comment is almost like a contract in the sense that you actually testify what the code is supposed to do. Waaaay to dangerous. Input --> magic --> output is advertisement since Steve jobs used it a little too much in order to not have to do tech talk when presenting the iPhone. "A browser opens" ... "it's magic"

    But on a moreseriouse note: I think deep neural nets are not very well understood and evolution is also not 100% understood (how DNA works for example).

    We just need to closely monitor the output of a system to estimate the risk. Desired output = good ai. Undesired output: trash

    Edit: not an (a)i but anyone afraid of trump having the nuke codes?
     
  14. Cornucopia

    Cornucopia Well-Known Member IQ: 140+

    One could argue that we don't understand the inner workings of anything completely. Given that our understanding of the smallest building blocks are limited.
     
  15. Cornucopia

    Cornucopia Well-Known Member IQ: 140+

    What is it about evolution that isn't fully understood? Mutations due to special circumstances?

    I don't think that the details are necessary to fully grasp the workings of evolution.
     
  16. conrad

    conrad Member Unclaimed IQ

    I've been working on team projects exclusively - so potentially permanent hell. :D. It's almost indispensible for programmers. Programmes who write code for themselves -or only the machine - won't survive long in a team, as least when it's 'professional'. One of the most important technology is source code management. "Code that isn't checked in doesn't exists. " lol

    But seriously, there are very good teams in which working is what ones imagines coming close to heaven. Colleagues do all this things: commenting, clear "self-speaking" code, comments on check-ins, comments on changes, well-organized inferfaces, communication among each other. Since 1995, when I started, these have always been the key aspects of good team-work in programming.

    These cases I've mentioned were two extremes. The first were 2 functions from an intern - and, really, without talking somebody down: the MOST terriblc I've ever seen. Total jumble. It worked, but nobody dared to change a single character. Credo for any fledgling programmer: you don't write code fo yourself and the machine, you write it for your colleagues. They're one who have to understand it. The machine understands it anyway.

    Impressions from the dungeons of programming.
     
    Last edited: Jun 15, 2017
    AguirresPriest likes this.
  17. conrad

    conrad Member Unclaimed IQ

    You seem to have at least some experience and some traumatic impressions in team-friendly (-hostile) programming with a team size of >=2. Good potential.
     
  18. conrad

    conrad Member Unclaimed IQ

    This is exaclty what I don't advice anybody to do. Many unexperienced programmers, who are coming right out of university, haven't learned yet how important comments are. They should not explain how the codes works (that's a matter of good coding), but what is does and how to invoke it, hinting to possible interconnections. Commenting source code is usually not taught at university, not even broached. The focus there is on algorithms, implementing them so that they work and produce the right output. The code snippets from university staff members are sometimes not most pracital ones (variable names like a, b,c .. for example).

    Young programmers are usually going through a tough school at their first job, and they're lucky if they have more experienced colleagues who can serve as positive models - those who are neither talking down to them neither badmouthing them, but help them to become better.
     
  19. Creedinger

    Creedinger Well-Known Member IQ: 120+

    Not commenting your code is totally stupid no matter at which level. It is an easy and powerful tool to understand your code better when coming back at it.

    Good variable names are important and should indicate the type and content at least.

    Same goes for maintaining a clean object oriented design using patterns where possible but with "keep it simple" in the back of your head.

    In a team work according to a development process e.g. Agile development.

    The part of learning a language on the first job makes comments even more important ;)

    I'm not a programmer anymore and all I do is write some scripts when necessary ( in a gosu language such as python ).

    Back to deep neural networks and how they are understood ;)
     

Share This Page