1. 28
  1.  

  2. 3

    Fun idea! I think it says more about wine culture than any amazing capability of Markov Chains.

    1. 3

      The descriptions are really just strings of adjectives, which is why this is a simple task for markov chains (having no real knowledge of sentence structure).

      If you used contradictory adjectives to describe a wine, you’d probably get called out.

      1. 3

        The flavour juxtaposition simply makes the wine better, not the reviewer worse!

    2. 3

      I like the wine critiques, but this mixup of HP Lovecraft and Puppet documentation is probably the best/funniest Markov text generation I’ve seen.

      1. 2

        Very cool! Seems a lot easier than I thought it would be (although I have yet to dive into the code).

        1. [Comment removed by author]

          1. 1

            If you are planning to fool an investor who doesn’t know much about wine or anything, then all you have to do is spend one day reading about wine and the investor wouldn’t be able to tell you apart from true ‘experts’.