Lattice
A growing lattice of mental models across disciplines — rooted in Charlie Munger's idea that genuine understanding comes from building a latticework of world models, not from one-dimensional expertise. Most of the entries below live under "Other" for now; I'm sorting them into Math, Engineering, Accounting, Bio, Physiology, and Psychology as I go.
18 entries
No. Term What it means Theme
01 Abundance → Scarcity When technology makes X abundant, a complementary Y becomes scarce — and that's where the monopoly forms. Other 02 Activation Energy The cultural leap from zero to one real deployed customer — the hardest, most catalytic step. Other 03 Asymptotically-Free Inference The cost of running a model trends toward zero — so the edge moves to whatever stays scarce when inference is free. Other 04 Combinatory Play Remix mental models across books and domains — with an LLM as cross-referencer — to find non-obvious connections. Other 05 Conjectures and Refutations Knowledge grows by bold conjectures exposed to hard refutation — the engine under iterative 'loops.' Other 06 Controlling the Tempo If delay is free, customers delay forever — so engineer a forcing function: a hard date and limited slots. Other 07 Credibly-Neutral Verification A 'Switzerland' third party that certifies agentic work with no vendor agenda — an Okta or Moody's for AI output. Other 08 Don't Ask, Tell Customers often can't spec their own needs — lead them with a framework instead of asking what they want. Other 09 Durability vs. Momentum Underwrite rounds on durability, not momentum — high durability paired with low momentum often means a cheaper entry. Other 10 IBM and the PC In a modular system, value concentrates in the layer that can't be swapped out — IBM ended up owning PC assembly (a commodity) while Intel (chip) and Microsoft (OS) owned the bottlenecks. Other 11 Ideal Customer Profile Narrow who you sell to first — a sharp ICP compounds; a broad one dilutes learning and proof. Other 12 Lighthouse Customers Structured early-customer relationships with explicit phases and success criteria — not casual pilots. Other 13 Objective Functions over Algorithms Increasingly you don't write the algorithm — you define the goal and let the system iterate toward it. Other 14 Ring the Bell Define exactly what pilot success is before it starts — then prove those precise criteria were hit. Other 15 Super Thinking & Super Building Two founder qualities to underwrite: super thinking (seeing ten steps ahead) and super building (throwing an idea away and rebuilding). Other 16 The Bitter Lesson General methods that scale with compute beat hand-coded human knowledge — over the medium term. Other 17 The OODA Loop Observe–Orient–Decide–Act: don't decide and act before you've genuinely observed and oriented. Other 18 The Shifting Bottleneck There's always a bottleneck; when technology moves it, build where it's about to land. Other