Well thats settled

4 Comments

  1. So unpredictable behavior = stupidity/insanity? I think he might be oversimplifying things a bit but not too much. I guess you’ll be taking some neuro-biology classes now?

  2. I have no personal stake in solving brains. My dataset (machines) is already solved; in fact, they usually come with instruction manuals.

    But he isn’t talking about the predictability of behavior, he is talking about the ability to make predictions. He argues that ‘making predictions’ is not a behavior at all, it is a memory prediction framework that is good at recognizing temporally distributed patterns. He believes this framework characterizes all intelligence, or at least the processing occurring in the neocortex.

    Read more here.

    There is a simple platform running this framework that is free to download for developers on the Numenta website, or you can see this quick video of the kind of output it produces.

    I need to actually read the book, but the Wiki article describes the actual neurobiological predictions made by this theory, which I found particularly impressive.

  3. Since his being right would be friendly to broad claims I want to make about normativity being built into our cognitive experience, I declare him to be right. Also, I predict that he’s right.

Submit a comment