Comments by "Taxtro" (@MrCmon113) on "Computerphile" channel.

  1. That makes me happy I started computer science in the summer, so our first language will be Haskell rather than Java.
    22
  2.  @cmonkey63  Machine learning describes precisely what it's about. Really, I cannot think of any better term for it. Computer aided reverse deduction? Knowledge discovery in databases? Automated stochastical analysis? Practical function fitting? Those are all obscurantist, learning is what it's about. And who learns? A machine.
    14
  3.  @kyrothegreatest2749  What is "sentient" supposed to mean?
    12
  4. Fortunately I swallowed my drink before reading that comment.
    8
  5. That is a description of a Turing machine. It's just unnecessarily complicated. The employees are the states and the split in the stack marks the position of the read-write head.
    6
  6. He's a professor? He looks so young - and where is his beard? My professors all have beards.
    6
  7. Can you link where you got that from?
    6
  8. alex7071 Understanding it << coming up with it yourself.
    6
  9. It should be obvious, but to very many people it is not. It's an ethical confusion.
    5
  10. Guilherme Moresco The death of those we idealize often hits us hardest. Certainly was the case with my professor recently.
    5
  11. Mohamed Al-Ganzoury It's called malicious compliance.
    4
  12. "Category" and "Functor".
    4
  13. Seegal Galguntijak You don't need a genius for that. Here BB(n). There is your mathematical formula. Thing is that you can write formulas for anything, but that doesn't make them computable. What you are looking for is an algorithm and because algorithms are just as powerful as TMs, you cannot construct a computable function growing faster than BB by definition.
    3
  14. Disthron A neuron is basically like a PLA. It's more complicated than a single transistor, but it works like a logic gate. It either fires or it doesn't. In that way the brain works with binary just like our computers.
    3
  15. ghosttwo2 Yeah weirdos like the inventor of gifs...
    3
  16. No, it's both. We knew about lots of the best machine learning algorithms more than thirty years ago, but we didn't have the datasets to train them sufficiently. Deep neural networks are comparatively simple, but they perform miracles if you throw tons and tons of data at them.
    3
  17. "Kind of impressive" is a massive understatement. It's one of the most awesome and scary things I know.
    3
  18. Physics Videos by Eugene Khutoryansky You don't understand motivations. Would you rewrite your brain not to value the well being of conscious creatures?
    2
  19. Of course. Recursive definitions are for a better understanding. In code however they may lead to a horrible runtime and memory usage.
    2
  20. What human? People disagree. Should it never do anything an Islamist disagrees with? An AGI needs superhuman miral judgement. A measure that is independent from any particular person.
    2
  21. He means the number that is printed on the tape, not the number of possible TMs (the one he showed the formula for).
    2
  22. Correct, quantum computers are exactly as powerful as turing machines or lambda calculus.
    2
  23.  @returnexitsuccess  How does the word "degree" suggest that it's not an absolute scale? "Degree Kelvin" sounds off, but I don't see how it's wrong.
    2
  24. How is that different from other languages and interpreters? I'm pretty sure you could do the same thing in C++ or Java or Haskell without any "duality". Also I thought computer scientists had largely abandoned meta programming.
    2
  25. Russels' paradox is about a set of axioms leading to a CONTRADICTION, not it being incomplete. The question whether the barber shaves himself is not undecidable, but both options are clearly wrong.
    2
  26.  @victorlevoso8984  Only if the if allows you to jump backwards in the program.
    2
  27. This works for arbitrary iterations of the word buffalo. Buffalo. Buffalo buffalo. Buffalo buffalo buffalo. ...
    2
  28. How has Thorsten been able to keep his accent for so long?
    2
  29. No. It's when you define something in terms of it's precursor and a base case.
    1
  30. Instead of pressing the key, you draw the sign.
    1
  31. Remi Caron Well, you would only want an AI, which values human well being. So, this super intelligence would be like an adult waking up in a group of toddlers in trouble. Of course the adult will impose his ideas and plans on the toddlers, but he'll do it for their sake.
    1
  32. Damaged Pylon Yes, but it doesn't care. I know that my sense of morality arose as an evolutionary advantage over competitors, but that doesn't decrease my compassion for animals, which will never thank me for helping them. You see how that works? It doesn't matter how your values arose.
    1
  33. Yeah, to me it's just a complete non-sequitur. If the basilisk already existed and hardcode a response into itself that would be another story.
    1
  34. The ants are encircling your foot asking themselves where you could possibly escape to.... The problem is not physically turning off the computer that runs the AGI, the problem is that you wouldn't even want it. You wouldn't know when to press stop. You'd defend that button with life. That's how intelligence works.
    1
  35. No, we need to make sure that the expected outcome is slightly better than status quo and then go for it.
    1
  36. It's fun. Welcome to mathematics.
    1
  37. That's word games. AI is extremely useful and already commonly applied. You can call it an algorithm, but so is the workings of your brain.
    1
  38. The well being of conscious creatures must be the number one objective of any general AI.
    1
  39. True. Other automata like pushdown automata are more relevant to actual applications. Turing machines are just a very simple way of describing something very powerful, but the way one would work on a specific task is unnecessarily complicated. There is no reason to mimic Turing machines in technology.
    1
  40. @Mswati Well that's about those specific problems then and not about recursion itself. You don't have to be smart to understand recursion or induction.
    1
  41. No. Those are all detrimental to your goals. If you hardcode stuff like that, you might as well make a classical algorithm. A set number of stamps is just another goal, which might be achieved in immoral ways.
    1
  42. UrbanTarzan Duh You are deeply confused. Humans are not motivated to spread at all. Also stop using the word "purpose". It's confusing.
    1
  43.  @KipColeman  No. That's like a dozen ants trying to encircle a human.
    1
  44. You obviously don't understand Gödel. Gödels incompleteness has nothing to do with learning. There is a similar circumstance for learning (no free lunch theorem), but theoretically there is really no problem with AGI. You know why? Because we exist. And AGI is no more than human level intelligence. Do you think Gödel proved that humans don't exist?
    1
  45. Exactly. The configuration of the automaton is different from it's state. At any moment the automaton has a configuration, and that consists of it's state and the position of the read/write head on the tape.
    1
  46. ARTUN3 No, that would be totally foolish. You want a GI to think about solutions, you couldn't think about. So your pathetic attempts to hardcode idiotic rules would destroy it's functionality. It's like a toddler making rules for it's parents.
    1
  47. That's the same as a finite tape.
    1
  48. The transition functions are not read; the Turing machine knows them from the start and all of the time. Nor is it important that the alphabet is 1s and 0s. You can expand the alphabet as you wish without the machine becoming any more powerful.
    1
  49. void* You are the most confused person, I've ever seen. Pick up any book on history, please.
    1
  50. I don't know what you mean with simulating memory. The Turing machine is essentially a function. You give it an input string and it may give you an output.
    1