A collection of some wisdom for working with models.

Some lessons may also apply to humans. 😉

Strategy

  1. Build systems that leverage models; don’t build models.
  2. Expect models to get smarter and cheaper.
  3. Create systems that perform better when scaled.
  4. The human brain is a beautiful starting point for agent design, but also build simple systems that scale with more compute.

Thinking Well

  1. Novelty requires some randomness.
  2. Models discern great ideas quicker than they create them.
  3. Big problems require lots of thought.
  4. The quality distribution of thought/experience is Gaussian.

Prompting

  1. Never encourage in prompt when you can enforce with code.
  2. Make the models’ job as easy as possible.
  3. Know when not to give the model data.
  4. If your prompt is long, you’ve done something wrong.