Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

    •  
      CommentAuthoraber0der
    • CommentTimeJul 15th 2021
     
    Posted By: AngusI have noticed that whenever I make a purchase over the interwebs I am subsequently deluged with advertisements for the very same thing that I have already bought. This does not strike me as intelligent.


    Advertising in general is not the industry/gewerbe(haha) to attract the brightest minds. Since their jobs are so boring they employ their creativity for mobbing.
    •  
      CommentAuthoraber0der
    • CommentTimeJul 20th 2021
     
    Tesla FSD beta 9 - some interesting and amusing views

    https://youtu.be/wD_mF0OLJPs
  1.  
    Nobody understands deep learning. Otherwise we'd just program the weights for the task at hand and we'd be done.
  2.  
    https://www.quantamagazine.org/how-computationally-complex-is-a-single-neuron-20210902/

    A real neuron is a lot more complex than an artificial one.
    • CommentAuthorAsterix
    • CommentTimeSep 3rd 2021
     
    I think the real complexity is in the connections between neurons.
    • CommentAuthorloreman
    • CommentTimeSep 3rd 2021
     
    The real complexity is in the tendency of space time to clump into stuff that is somehow different from other stuff
    •  
      CommentAuthorAngus
    • CommentTimeSep 3rd 2021
     
    Deep. Headed for the Moletrap Dialect.
    • CommentAuthorloreman
    • CommentTimeSep 3rd 2021 edited
     
    Thanks. I’ve just had a very strong Maccas hot chocolate and it’s brought on a bout of profundity.
    •  
      CommentAuthorAngus
    • CommentTimeSep 3rd 2021
     
    There there. Just have a little lie down and it will probably subside. Although I'm not sure if the subsidence of a bout of profundity is a good idea, now that I think on it...
    • CommentAuthorloreman
    • CommentTimeSep 3rd 2021
     
    I’m not sure that buying anything from Maccas is a good idea either (but I did win a free McChicken when I peeled the sticker off the side of the take-away cup-Oh happy days!)
  3.  
    I wish you the joy of the free McChicken
  4.  
    •  
      CommentAuthoraber0der
    • CommentTimeSep 11th 2021
     
    I like that, but I also don't.

    :|
    • CommentAuthorkorkskrew
    • CommentTimeSep 11th 2021
     
    I've always thought that Reality is Irrelevent would be a better name for the Two Minute Papers channel.
  5.  
    Replacing the backpropagation algorithm with something more physically realisable in an organic system has long been one of the goals of neural network research. Several years ago I posted here how impressed I was with the work of Blake Richards and the YT vids of his lectures on possible learning mechanisms for the pyramidal neuron. It's good to see that coming of age here.
    https://www.quantamagazine.org/brain-bursts-can-mimic-famous-ai-learning-strategy-20211018/

    Rivalling the efficiency of brain learning with synthetic systems is not a problem that we can expect to solve overnight. Just to mention two aspects: the one-shot learning trick, and the tiny power consumption per node trick.
  6.  
    https://www.youtube.com/watch?v=psTb_Voxpqc

    GLOM - Son of Hinton's capsules, also by Hinton.
    •  
      CommentAuthorTrim
    • CommentTimeDec 5th 2021 edited
     
    As Venus is lacking in water but has plenty of carbon, silica and energy why not build large AI's there?

    Starting in the cooler atmosphere supported by large nitrogen and oxygen balloons and rings, initially helped a lot by humans and heat tolerate robots it will teach itself to 'thrive' on the hot surface of the planet where it will learn to replicate.

    Programmed with the so called 'Venus project' it's vast brains would be happy to help humans solve their problems and it wouldn't need money in payment and no logical reason to harm humans.
    •  
      CommentAuthorAngus
    • CommentTimeDec 5th 2021
     
    Bonkers.
    If it's Venus hot everywhere your "silica" AIs won't work, even if they are really made of silicon.
  7.  
    What size are they?
    • CommentAuthorBigOilRep
    • CommentTimeDec 5th 2021
     
    Posted By: AngusBonkers.

    You don't say...