North-Star Thesis

In one sentence, in Jacob's own words:

"Now, what I'm actually building is a massive-scale collective cognition system for humanity. But one of the pieces it takes is a personal knowledge management system." — 2024-03-22, #explainingideaflow #best #pitch

That's the whole thing. Every other framing in this wiki is a projection of that statement onto a specific audience or a specific layer.

The three-scale decomposition

IdeaFlow is building one substrate that works at three scales. Same data model, same interaction grammar, same graph.

Personal — a framings/graph-database-for-your-brain — a framings/notebook-for-our-minds — "quick capture, quick sort".
Team/company — a collective brain — "the ultimate CEO dashboard" — a Slack-listening bot that harvests unstructured conversation into structured knowledge.
Humanity — a framings/neocortex-for-the-global-brain — the substrate under framings/team-humanity coordinating at civilization scale.

See system-map for how these stack.

Why "cognition" and not "notes"

Jacob's repeated insistence, across years:

"Reducing intelligence to generation is like reducing software to print() statements." — 2024-02-15

The rejected framing: notes app. The insistence: cognition — capture, cluster, relate, surface, augment, coordinate. Notes are the raw material; cognition is what the substrate does with them.

Cody Hergenroeder's shorthand for this at the product level:

"O(1) capture and recall is how I've been phrasing it to myself."

Why a graph, not a document

Jacob, citing Taylor Mitchell's early-2024 framing:

"In most outliners, when you 'expand' a node, you're showing its children. This is different from graph database explorers, where when you 'expand' a node, you're showing all related nodes, not just children."

A graph is the right shape because the same spark (framings/sparks-of-motivation) appears in many contexts at once — planning, health, work, relationships. A document forces a single ordering that hides everything else. See also framings/graph-database-for-your-brain.

Why now

The markup burden that killed the 2000s Semantic Web is gone:

#ExplainingIdeaFlow #whyNow #SemanticWeb — LLMs can now produce structured graph nodes from unstructured input.

Jacob's collaborator context: Tim Berners-Lee was one of Jacob's research advisors at Oxford/MIT (see vision-convo → entities/tim-berners-lee). The lineage runs straight from the Semantic Web through the LLM era into IdeaFlow.

The long bet (civilization scale)

From vision-convo's Vision for the World theme:

"We are moving one way and another part of us moves the other way. We're fragmented … We want to be the fish, not the cat-dog."

IdeaFlow is the substrate on top of which civilization can become the schooling-fish body instead of the cat-dog. Concrete applications: manifestos.world, World Issue Tracker, World Progress Bar, Accretive Collective Action.

system-map — how the three scales stack
what-it-is-not — category-confusion prevention
chronology — how this thesis was refined 2022→2026
canonical-terms — the working vocabulary
pitches/best-pitch — Jacob's own best compression of this thesis