The Architecture of Learning: From Statistics to Intelligence(little-book-of.github.io)

75 points by scapbi 5 days ago | 8 comments

  • thorum 1 day ago
    "not a X, but a Y" - 8 matches

    "is more than a X... it is Y" - 3 matches

    "not just X, but Y" - 4 matches

    [-]
    • jal278 1 day ago
      Yeah -- I don't get why this is front-page -- reads like LLM quasi-insight:

      "Through activation, lifeless equations became living systems. The neuron was no longer a mere calculator; it was a decider - a locus of transformation where signal met significance." -- wtf

    • godelski 1 day ago
      More than that, it uses a lot of words to say so little.

      Even a quick scan shows some pretty critical errors. In 76.4

        Two parameters govern its perception:

   • ( ): neighborhood radius
   • ( MinPts ): minimum points per dense region
      Or later in 76.7

        In Fuzzy C-Means (FCM), each point (x_i) receives membership values (u_{ik}) in (
                               0,1
  ), satisfying (k u{ik} = 1). The objective is to minimize:
      These are not human mistakes. They are categorically different

      Also, the math really smells of AI. It has equations but it is like they have no substance. It has the form, but not the feeling. I know all this math and looking through I don't know how anyone could learn from such text. I'm not sure how it could even serve as a good reference. Where are the derivations? Where are the corollaries? Where are the implications? The extensions? The... depth?

      0/10. I think you would be worse off by reading this

    • ForceBru 1 day ago
      Why not just say what you want to say??? Surely these statistics are supposed to suggest some "obvious" conclusion, probably that the article is somehow bad. What do you mean by these numbers???
      [-]
      • SchizoMode 1 day ago
        They're saying it was written by an LLM because of the style of writing.
  • pfekin_2nd 1 day ago
    This is brilliant, thanks.
  • bulla 1 day ago
    This is superbly written.
    [-]