Conversation

The more VUCA (volatility, uncertainty, complexity, ambiguity) in the world, the less useful the state-propagation model of foresight. State-propagation model: define boundary of 'important', book-keep what is in/out, propagate the 'in' bits via some model, ignore 'out' bits.
3
22
...the problem with the state propagation model under VUCA is that the state churns so much from in/out instability that there's no point forecasting more than 1 step out. Less obviously, memories become useless more than 1 step back. So you're temporal horizons shrink
3
7
...the alternative I've been increasingly using isn't as familiar because it doesn't have an obvious mathematical analogy in quantitative models. The key is to decouple assessments of what is important from mechanisms of propagation.
Replying to
... you decide what's important overall, in a cradle to grave sense, and track it through peaks and troughs of situational importance, whether or not it's doing anything interesting at a given time
4
4
...the situational foresight problem then turns into a 2-step problem: detect "interaction bundles" of individual important dynamics, and propagate past mixing events (kinda like billiard ball collisions). This is like tracking a bunch of interacting subplots of a story.
1
5