Synthetic A Priori

Seed idea: Can modeling a design project as a network provide underpinnings to better estimate risk and sequence problem solving? With connections to multi-scale systems, Alexander's idea of an unfolding process, and thin vs. fat-tailed variables.

Show Notes

Part 1: 00:12
Network theory, multi-scale networks, example of building a clustering feature on a side project, two scales of design: the feature level and the implementation level

Mentioned:

Part 2: 13:54
Unfolding as a network dynamic, learning at the fine scale under constraints from the large scale

Mentioned:

Part 3: 20:05
Risk, thin-tailed vs. fat-tailed variables, how underlying structure gives rise to different shapes of distributions, orthogonality and interdependence

Mentioned:
  • Nassim Taleb. See Probability, Risk, and Extremes (PDF) on thin vs fat-tailed variables. Re: the relationship between distributions and  underlying structure: "... we cannot rule out that it is not fat tailed unless we understand the process." (emphasis added)
Part 4: 30:14
Patchiness of risk in a design problem, observation in science vs. active control in design, targeting unknowns, redesigning at the feature scale based on information from the implementation scale, example of designing for independence

Part 5: 38:35
Scopes in Shape Up as tangled network neighborhoods, structure vs. opacity in the network, alternation between identifying structure and removing "the fog"

Mentioned:

Outro: 46:05
Is this thing on? Say hello or just raise your hand with a DM to @rjs on Twitter.

What is Synthetic A Priori?

Step into the messy phase of wrestling with ideas before they become conclusions. Ryan Singer draws connections between design, tech, science, and formal systems.