
    scalePattern = slow 16 ""


    do
cps (86/60/2)
d9 $ midicmd "midiClock*48" # s "midi"


    generateMelodicSeed = slow 4
  $ linger 0.5
  $ repeatCycles 3
  -- $ palindrome \n
  $ (+ (slow (irand (4)+1) (sometimes (inversion) (run (irand (4)+1)))))
  $ slow ((irand 3) + 1)
  $ e ("x"<~>(irand 8)) 8
  $ "x*16"<~>(irand 5)


    do
d9 $ midicmd "stop" # s "midi"
hush

index > /home/xinniw/Documents/garden/Mortal Computation.md

Mortal Computation

Mortal Computation is the idea that, in an analog computer, the computation dies with the hardware. I first heard this described in a paper about the forward forward machine learning algorithm by Geoff Hinton. He claims his FF algorithm could be implemented directly on silicon, with inputs represented as voltages and weights represented as conductances. This would be more energy efficient than simulating this circuit's behavior on a general purpose computer or doing a lot of ADC DAC conversion. He also claims that this chip would be more energy efficient to manufacture.

index > /home/xinniw/Documents/garden/Mortal Computation.md