The Iceberg

MATRIX observes effects → FRAMES detects patterns → PILLS infers mechanisms → DEEP validates causality → SOURCE reveals paradigms

The iceberg is an inference engine. Ten percent of it is observable from the surface — the events, the data, the effects. Ninety percent of it is below the line: the patterns, the mechanisms, the paradigms that produce what we see. This is the observation deck. The PG box is where you stand to look at reality.

Why an iceberg?

Because the interesting part of any phenomenon is the part you cannot see. A stock crash is an effect — its cause lives a kilometer down, in the paradigm that allowed the leverage. An AI model's output is an effect — its cause lives in the training data, which is a culture. Standing on the observation line, you only ever see the tip. The descent is the method for seeing the rest.

The layers are not levels of importance — they are levels of inference. MATRIX (0m) does not explain; it presents. FRAMES (−50m) detect patterns: taxonomies, timelines, maps. PILLS (−200m) infer mechanisms — and pills are hypotheses, never facts. DEEP (−500m) validates causality: consensus built or broken. SOURCE (−1000m) is the paradigm — bedrock. Ours is self-hosted infrastructure, trading algorithms, and the methodology itself.

"WE ARE COOKING KNOWLEDGE" — the source is the stove, the deck is the kitchen, and every fish is a dish still on the fire.

The descent, layer by layer

0m

The Matrix of Observation

Does not explain. It presents.

The surface where effects are first seen. Raw observation, data points, events — visible to anyone standing on the deck. Nothing here is interpreted yet.

MATRIX observes effects → FRAMES detects patterns

100%
-50m

The Frames of Pattern

Taxonomies, timelines, maps.

Pattern recognition. The layer that groups observations into structures: categories, sequences, maps. A frame is a way of seeing — useful, never final.

FRAMES detects patterns → PILLS infers mechanisms

60%
-200m

The Pills of Insight

Pills are hypotheses, not facts.

Inference. Heuristics and working models distilled from patterns. A pill is a lens you swallow and test against reality — it can be verified, and it can be wrong.

PILLS infers mechanisms → DEEP validates causality

25%
-500m

The Deep of Analysis

Consensus built — or broken.

Causal analysis. The fishbone goes tail to head here: effects traced to variables, variables to mechanisms, mechanisms to paradigms. The deep is where consensus is tested.

DEEP validates causality → SOURCE reveals paradigms

10%
-1000m

The Source of Paradigm

Bedrock. The paradigm itself.

The bedrock of the system: self-hosted infrastructure, trading algorithms, and the methodology itself. Paradigms are rarely seen and almost never moved — but everything above rests on them.

feedback ascends — SOURCE shapes every layer above

2%

The feedback loop

The flow is downward — observation to paradigm — but the influence is upward. A paradigm at −1000m shapes what the matrix at 0m can even observe. The deck is where the two meet: you stand at the surface, looking down the descent, watching the source steer the surface. SELF.HOSTED // NO.DEPENDENCIES