[cmax]

One idea I employ often is to "increase the temperature" of the system, non-essential complexity, bureaucracy, excessive variation of materials, etc becomes more apparent through a lens of being more demanding than necessary so that the physical limits of the problem dominates the solution.

Another idea I use is to mentally time travel and try to visualize how context-sensitive my decision/design/process is. Decisions that require a "superior" (or even accurate for that matter) understanding of the context I'd frown upon. Even if they are not mistaken and biased, to me, they promote a dynamic of continued survival through non-obvious and increasingly complex actions, rather than forcing a simple and obvious environment.

Another is engineering with a bias for optimizing recovery first then reliability (sort of a minimax).

So for example combined I would rather have a way to recover from broken code and produce a teaching moment than I'd try to prevent a developer from merging said code in the first place.

I'd try to exercise a process to recover all data from backups in a timeline that the company would survive first rather than dedicate resources first for redundancy and leave said scenario for later due to its low probability.

Take all of this in a "while there is value on the other thing, I value this more" wrap.

I like models that would force the space to be simpler and more obvious.