I’m going to go out there and be ridiculous and see what happens. I’ve got nothing to lose.
Wayde Van Niekerk
In building software, many things may seem unimportant but later turn out to be critical. Alternately, many things seem important that could be a false path. One heuristic that I have found useful is to think of things in terms of a feedback loop. At a task, is the thing you are doing speeding up the feedback loop(s) or blocking it?
The Spartan AI lifecycle is the spirit of this heuristic. Either you are speeding things up and making them more efficient, or you are not. If one inner ring of feedback loop needs to be optimized, say getting better unit tests or more reliable extract, transform, and load (ETL) service, ...