Three apparently independent objectives converge on the same direction. Recognizing the convergence gives meta-direction to the endeavor’s primary goal.
The three trajectories
1. Inference → determinism (skill maturity)
The skill maturity concept describes how skills move from inference-heavy (a capable agent reads instructions and exercises judgment) to procedural (a deterministic script encodes the exact steps) to tool (an MCP endpoint callable by any agent). Each promotion makes the skill cheaper, faster, more reliable, and more composable.
This is policy 001 (progressive automation) operationalized.
2. Specifying the endeavor (goal 006)
The primary goal is to fully specify the emsemioverse endeavor: what an endeavor is (semiotic-endeavor), what an implementation must do (semiotic-endeavor-specification), and how each aspect of the method fits into the whole. Specifications make the system measurable — you can check conformance, identify gaps, and improve systematically.
This is the content trajectory: knowledge moves from implicit practice to explicit specification.
3. Reducing LLM token usage
As skills mature and specifications stabilize, less inference is needed per unit of work. A procedural script uses zero tokens. A delegable skill uses tokens from a small, cheap local model. Only genuinely novel work requires a high-capability agent’s full attention.
This is the economic trajectory: the system becomes cheaper to operate as it matures.
The convergence
All three trajectories point the same direction:
implicit → explicit → measured → automated
- Skill maturity: inference → structured → delegable → procedural → tool
- Specification: practice → method → specification → conformance check
- Token usage: high → medium → low → zero
They are not three separate efforts. They are the same effort seen from three angles: operational (what agents do), epistemic (what the system knows about itself), and economic (what the system costs).
Meta-direction
Goal 006 (specify the endeavor) gives meta-direction to plans: it tells you which plans are highest-leverage by answering “does this plan move us closer to a fully specified endeavor?”
The convergence of these three trajectories gives meta-direction to goal 006 itself. It tells you which specifications to prioritize by answering: “does specifying this move work from inference toward determinism? Does it reduce the tokens needed for the next agent to do this work?”
Specifications that enable automation are higher leverage than specifications that remain dead letter. A semiotic-triage spec that enables scripted enrichment (replacing inference-based enrichment) is higher leverage than a semiotic-triage spec that just documents what we do.
This is the direction: every specification should be written with the question “what does this enable me to automate?” as a first-class concern. The purpose of a specification is not just documentation — it is the precondition for the next promotion on the skill maturity axis.
Relationship to existing concepts
- Closure pressure (ASR concept): the formal version of this convergence — skills under repeated execution pressure tend toward closure (deterministic, idempotent operations)
- Progressive formalization (agent-skill-interaction theory): knowledge moves from conversation → specification → skill → script → formal proof
- Operational closure (ASR theory): satisfaction relation connects the predicate graph to formal semantics; each layer of checking replaces inference with computation
- Escalation of communication (emsenn, 2019): ideas move from private to public venues, gaining rigor at each stage — parallel to knowledge moving from esoteric (fragment-internal) to exoteric (cross-fragment)