The shallow reality of 'deep learning theory'

Classical learning theory makes the wrong assumptions, takes the wrong limits, uses the wrong metrics, and aims for the wrong objectives.

In this sequence, I review the current state of learning theory and the many ways in which it is broken. I argue that the field as it currently stands is profoundly useless, and that developing a useful theory of deep learning will require turning elsewhere, likely to something that builds on singular learning theory.

Atomic Workflows

Clear's insight with 2 Areas/3 Notes/3 Sciences/9 Psychology & Psychiatry/Atomic Habits also offers a solution to orchestrating collections of habits in workflows. Let's call this approach "atomic workflows".

The process looks something like this:

  1. Enumerate all the habits that compose a workflow.
  2. Identify the overarching purpose of that workflow.
  3. Reduce to the minimum set of habits that achieves that purpose.
  4. Reduce those habits to the minimum set of actions that accomplish each habit's subgoal.
  5. Progressively expand these habits. (I.e.: 2 Areas/3 Notes/3 Sciences/9 Psychology & Psychiatry/Atomic Habits).
  6. Introduce new habits iteratively until you've reached the full workflow.