a complete graph has (n²-n)/2 edges for n nodes, so it's not quite the square, but it's proportional
(take away the /2 for directed graphs)
Conversation
anyway thinking about this from hamming's book, and…
Quote Tweet
2
1
…if concepts are nodes and analogical relations between concepts are edges, then you'd have a complete graph if all concepts related to all other concepts in a meaningful way…
1
…the complete scenario may not happen but consider a digression or parenthetical is a set of remarks on a concept in the context of all the other concepts you're writing about
1
anyway tl;dr the task in front of you could very well be proportional to x² of what you think it is instead of a·x
1
Quote Tweet
Had a thought that sqrt(n) to n^2 via n is a vastly more important phenomenon than 0 to 1.
sqrt(n) = random walk
n = muddling through (filtered random walk)
n^2 = network effect
This is what things like product-market fit really are. twitter.com/vgr/status/135…
Show this thread
1
2
Replying to
😎
Quote Tweet
Ideal subjective life distance equation
dₛ(n) = α √n+ 𝛿 < βn
Ideal objective life distance equation
dₒ(n) = 𝛾 n ²
near-random walk from the inside, network effect from the outside
2
2
Replying to
i feel like there is some graph theory/combinatorics that explains this that i am ignorant about
1
Replying to
Possibly this. It’s how I’ve always connected graph growth to compound interest. Network effects are the compound interest effects of preferential attachment dynamics. For an individual, just building on your own past work. en.wikipedia.org/wiki/Preferent
1
1
Replying to
now thinking out what the topological connection is, like activity on the surface of a sphere
1


