• tags: #complexity #ritual #behaviour #patterns #talk #dave_snowden #computing

• via future thinkers discord

• youtube.com/watch?v=l4-vpegxYPg

• references also: en.wikipedia.org/wiki/Cynefin_framework

• winnie pooh stories in britain, before disney

• reminds me of winnie poh characters as personality archetypes

• sense-making and sensemaking

• book: goodreads.com/en/book/show/818257 Karl Weik - Sensemaking in Organizations

• his (dave snowden) and kleins approach comming from a natural science background opposed to Karl Weiks

• topologies and taxonomies

• maps, models and frameworks

• (in IT) we need more chefs rather then recipe users , ( chef: some theory knowledge some practical knowledge)

• serependipity

• reality: we only see what we expect

• exaptation

• IBM, Microsoft, Apple exaptation

• how to design for serependipity?

• ecology vs engineering diagram

• orderd systems, complex systems, chaotic systems

• ~ 20 min ritual: "aligns the individual to the role that they are occupied on"

• "you don't see what you don't expect to see is at best you're scanning three percent of what's in front of you and matching it

against dominant patterns based on recent experience"

• "what ritual does is to change those patterns so you see the world differently"

• alcohol ban story

• archtectures and users finding out what they want

• socratique asking of questions

• pizza and coke

• description/stories of IBM bureaucracy - pizza and coke, tipping taxi

• "people will work around the system they won't follow the pattern if it doesn't work for them"

• "you wouldn't manage your children's party with key objectives targets mission goals project plans and milestones because everything would go wrong you can see a more elaborate version of that on the web

• definition complex system: if the evidence of a system supports conflicting hypotheses and contradiction can not be resolved within available time"

• "you can't understand the complex adaptive system through modeling or analysis you can only understand it by interaction with real-time feedback loops over multiple agents so you don't get cognitive bias"

• "if you have a complex system and you intervene the only thing I can guarantee you is that you will have unintended consequences" frog australia example

• "..that means you are ethically responsible for them(consequences), big new thing for the government"

• need many small experiments in parallel to filter out negative ones

• ~ 35 min explaining the cynefin framework for the complex domain : DS its important to test in parallel (concurrent?) not in sequence

• the lean startup book didn't look at the companies who did fail

• liminality antropology

• 1 year sprints

• new approach in the complex domain: cluster matching of unarticulated needs vs. frustration narratives

• ~ 49 min "reasons is we're reducing human beings to information retrieval and process and AI is better at that"

• the power and major importance of context: ".. and by the way you can write them 10% of what you know so anything which is processing text is again dealing with a very very limited set"

• shifting to the adjacent possible within context

• architecting for discovery instead of architecting for delivery

• fractal change, multi-perspective maps

• scaling in complexity: not by aggregation or imitation, but by decomposition and recombination

I generally agree with the approach of building (large) programs by creating a set of DSLs that make the different parts of the program easier to express. This is one of the reasons I prefer languages that allow user extension (like Lisps). Glad to see the author include a link to Racket in the footnotes, as their Language Oriented Programming research agenda is very explicitly about this sort of thing.

Wanted: a live coding system with Racket-level support for little languages and Smalltalk-level support for tooling.

FYI - I’ve switched from writing Lisp macros using Trees (AKA Lists) to writing macros using Characters with Ohm-JS.

Any folks here use Raku? Started poking around its docs over the weekend and was pleasantly surprised to see it supports the JVM