17:41
Watching this MIT/SDM talk by Richard de Neufville {{video https://www.youtube.com/watch?v=A2-g_ybdnzA}}
what I think he’s describing is what I might call Bayesian design
which would involve combining System Design with proper Bayesian updating
thinking of opportunity cost
in other words, designing for the distribution of possible outcomes
understanding that losses and gains are not always symmetric
The world of coming up with general ideas for phenomena across different sectors, or Abstract, is a really messy one
Hard to know when you’re discovering something real vs when you’re stumbling onto some personal mental pattern that’s mostly BS (ie either useless or arbitrary)