Everything by default diffuses.
This is the second law of thermodynamics: entropy.
Diffusion creates variability / diversity, which is the necessary ingredient of innovation (more likely to have exposure to what turns out to be the right idea).
But to create something that coheres and endures, it often requires coordinating multiple people at the same time to build it.
One way to do this is focus: "this is the thing we are doing right now, ignore everything else".
Focus can help build things (and ultimately nothing matters if it doesn't happen in the real world); but it reduces diversity and thus resilience.
What if the right idea to survive or thrive is not in the region you're focusing on?
Given this, how can you make good things accumulate in the real world?
There are two complementary approaches: breadthwise and depthwise.
Breadthwise would be configuring it so the various actors don't have to coordinate to accrete things of value.
The non-coordinating group of actors is a swarm.
This is one of the benefits of ecosystems; the swarm finds and builds many random things, and the good things hang around while the bad things erode away.
The overall cost of the effort of the swarm is massive, but for any single participant (especially one who benefits no matter who in the swarm finds it) it's minimal.
Depthwise would be slicing your actions into smaller actions.
Instead of doing one big thing that takes a lot of time, do it as a series of smaller steps, where you can adapt in between steps.
This helps your OODA loop tighten, giving you diversity/adaptability while allowing instantaneous focus.
Of course, this isn't a panacea:
Sometimes the administrative burden of many small pebbles is much higher than one big boulder.
The process of sharing information / deciding on the next step (coordination) has super-linear costs.
So sometimes it's best to pay the coordination cost once for a big boulder, but at the cost of adaptability / resilience.
This "small number of pivots" logic is the "big rig" style of organizations.