Reading this for the first time, ten years after publication, it’s a mix of comfortingly out-of-date technologies and amazingly ahead-of-their-time thoughts (eg, peer-to-peer file sharing among private groups). Covers a vast amount of ground, from ecosystems to 3D graphics to evolution. I so wish I’d read it in 1994.
9 Bees - election hall for idiots - true nature of democracy. Emergence.
28-30 Benefits of swarm systems: Evolvable, resilient, boundless, novelty. Disadvantages of swarm systems: Noncontrollable, nonpredictable, nonunderstandable, nonimmediate.
31 Hardest lesson to learn: “organic complexity will entail organic time.”
32-33 Past: atom. Future: net.
35 Braess’s Paradox: Adding capacity to a crowded network (eg, streets) reduces overall production (eg, more crowding).
58 Use of “peer to peer”.
58 “Complexity must be grown from simple systems that already work.”
80 Ecosystems. Sensitive to initial conditions. Stabilise to different states depending on the order in which species are added.
85 More apparatus is necessary to evolve intelligence (eg, thumb) than to sustain it.
103-6 Rocks, raindrops, etc. are made of once-living matter.
111 Four types of game: Chicken, Stag Hunt, Deadlock, Prisoners’ Dilemma.
114 If you stick to any strategy in a changing world, it will be out-evolved eventually. Randomness creates long-term stability.
114-5 Zero sum and non-zero sum games.
116 Gorbachev’s co-evolutionary cold war strategy. In non-zero sum games it may make sense to announce your strategy so others can adapt to it.
135 Complex systems die creatively.
148 The only bearing in biology: at the joint of a sperm’s spinning hair propeller.
149 The industrial revolution’s self-governing steam power was the first phase of the information revolution.
157 Command economy vs price regulation. Calculation vs cybernetics.
161-2 Three stages of the advent of automatic control: Control of energy (steam); Control of materials (informing matter with many feedback loops, eg, microchips); Control of information (we need to harness the explosion of information).
163 “The most intelligent control methods will appear as uncontrol methods.”
167 San Francisco Aquarium’s coral reef environment survives on electricity - no external food.
181 A team of eight people is ideal for any complex, hazardous project.
195 Toxins taking months to work through body. Emerging from fat reserves.
221 Adaptive office environment.
223 Local knowledge often better than other experts.
224 “Adaptable technology means technology that will adapt locally.” Local knowledge is required to operate VCRs and mobiles.
225 Danny Hillis on the benefit of a standard, not customisable, interface: “The reason we create artificial environments instead of accepting natural ones is that we like our environments to be constant and predictable. We used to have a computer editor that let everyone have a different interface. So we all did. Then we discovered it was a bad idea because we couldn’t use each other’s terminals. So we went back to the old way: a shared interface, a common culture. That’s part of what brings us together as humans.”
239 Cyberspace is a resource that increases the more it is used.
249-50 Adaptive technologies.
251-2 Two types of complex systems: continuous (car-handling) and discontinuous (complex software).
254 “Poka-yoke”: A Japanese invention for error prevention. eg, “A holding tray [on an assembly line] may have a specific hole for every bolt so that if there are any bolts left the operator knows he missed one.”
257 Herbert Simon - “satisficing” - good enough.
260 Teilhard de Chardin - wrote about hardwiring the collective consciousness.
269-70 Group file-sharing in secret.
303 Sim City etc as adaptive technologies.
322 Keep adding more to the Net and it becomes something different.
373 Darwin Chip - evolving software and adapting to work patterns.
439 Difference between adaptation and evolution: Adaptation bends a structure to fit a new hole. Evolution reshapes the structure itself, how it can bend, creating new holes.
441-2 Evolution breaks down barriers, opening new spaces. Unexpected. Not just creating variants.
450 Artificial life: a few lines of code that takes years to run.
454-5 Evolution is not a synonym for change.
460 Benefits of genetic adaptation.
464-5 Cultural vs genetic change - causes and effects.
470 on. Postdarwinism - Lamarckian feedback.
489 Richard Goldschmidt - Hopeful Monsters - small changes in embryos can create large changes in final form. Evolution can “jump” - not solely transitional.
490 This explains gaps in fossil records.
491 Species persist - stasis - for a long time, then change/branch suddenly. Complex systems cannot change gradually and still function. [Like ideas? Paradigm shifts?]
513 Sparse networks don’t adapt well to change. Denser networks do. But very dense networks don’t.
516 There is a sweet spot between a frozen repetitious state and a “noise” state. As a system approaches it, it tries to stay there - “surfing the wave”. Systems that are most adaptive are so loose thy are almost out of control. “Life”.
517 Rigid systems can improve by loosening up. Chaotic systems can improve by getting more organised. The universe - if a few variables were slightly different (gravity, mass of an electron, etc.) the universe wouldn’t be anything like it is. No life?
524 A theoretical model of evolution: Globe, expanding over time. Each species at any point in time occupies one point on the surface. Therefore each species is equally evolved - has spent as much time evolving as any other. (Also the space of possible evolved positions - the sphere’s surface - is always expanding.)
537 A mutation rate of about 1 in 1000 is average.
542 Complex adaptive systems need to anticipate.
548 1: You can make predictions from chaos’s under-lying patterns. 2: You don’t need to look far ahead for it to be useful. 3: Even a little bit of information about the future is useful.
562-3 Theodore Modis’s three things that help us make long-term predictions: Invariants (some factors change little); S-curves; Cyclic Waves.
572-4 Critique of Limits to Growth model: Narrow overall scenarios; wrong assumptions; doesn’t take human learning into account; no geographical differences; inability to model open-ended growth.
577 It can stabilise, but unlike living systems it can’t learn, grow or diversify. Therefore it will fall behind reality.
578 It requires outside control for adjustment. “Vivisystems” (eg, economies, countries, human culture) can’t be controlled externally.