1. 22
  1.  

  2. 5

    I am disappointed at the title of this blog post, because it insufficiently captures the sheer scope and importance of the topic discussed within.

    1. 4

      It is an excellent post. It packs a lot of content into a small word count with clear explanations.

      1. 4

        thank you for convincing me to read to article instead of just clicking through the comments! I agree it’s an excellent post :)

        1. 1

          Agree. Wasn’t expecting it to be that good. Updating my philosophy to give friction a better reputation.

      2. 2

        Same, there’s just so much potential to this article

      3. 4

        I don’t think any of the examples listed in this article are examples of this law. They’re examples of the state of things being different to the author’s ideal state of things, but they’re mostly negative side-effects of complex and multivariate undirected processes, not targeted and optimised-for goals.

        1. 1

          Obliterate all privacy, and make all the information about all people, governments, and other organizations available to everyone all the time

          Just as a thought experiment, Rather than mandating maximum data collection, simply don’t. I think the effect would be that people would be much more careful about which data they collect. For example if a social media platform started collecting way too much personal data about customers, customers would be aware of this and seek out a competitor, creating a pressure to reduce the data collected. If you could access every webcam in every home over the internet, sales for cameras with mechanical shutters would skyrocket.

          I am not saying this would not be a dystopia, merely suggesting that the suggestion of putting cameras in every room makes it far worse and is not necessary. Just making public whatever data gets collected still hugely increases the capacity of the model, forcing more data to be collected feels like overfitting again.