Some guy.

  • 1 Post
  • 43 Comments
Joined 1 year ago
cake
Cake day: June 11th, 2023

help-circle















  • Yeah, a flair or a robot is way quicker than a powered espresso machine.

    I’ve been trying to optimise my workflow using a stopwatch and doing as much in parallel as possible. The key is to have water boiling and beans grinding simultaneously, and then milk heating and espresso extraction simultaneously.

    I can make a flat white and be all cleaned up and packed away withing 4mins.

    Process:

    • add water to kettle and start boiling
    • add beans to grinder and start grinding
    • get robot off shelf and put into position with scale
    • put milk into French press and in the microwave with time set to 1min (but not yet started)
    • grinding has now finished. WDT and tamp.
    • kettle has now boiled, press start on microwave
    • water into portafilter and press (~30sec)
    • empty and clean portafilter
    • remove milk from microwave, froth, pour.
    • put away robot, clean french press.




  • It’s the defacto term for how we fit a statistical model to data, unrelated to any copyright concepts. I’m pretty sure we called it “training” back in 1997 when I was doing neural networks at uni, and it’s probably been used well before then too.

    Neural nets are based on the concept of Hebbian learning (from the 1930s), because they are trying to mimic how a biological neural network learns.

    This concept of training/learning has persisted because it’s a good analogy of what we are trying to do with these statistical models, even if they aren’t strictly neural networks.