⟨ H, D ⟩ durable artifact
Authoring a Z₁ artifact
A human writing code, a shell pipeline, a SQL query, a regex. The authoring act is paid once. Subsequent executions run at the substrate’s thermodynamic floor.
The 2 × 2
Four cells.
An agent’s location reduces, in the simplest model, to two values: who occupies the seat, and how they select. The other two axes (primitive, timestamp) are inherited and universal — not classifiers. The classification of agents reduces to the 2 × 2 of (Actor × Selection).
The four cells
Where every agent currently lives.
σ = D · deterministic
σ = S · stochastic
A = H · human
⟨ H, S ⟩ source of novelty
Deliberation
Wetware noise drives selection. Judgment, taste, design, prose. The selection cannot be reproduced — not even by the same agent under the same conditions.
A = S · software
⟨ S, D ⟩ durable artifact
Compile, don’t loop
A compiler, a code generator, a scheduler, a language-model runtime in author mode. One stochastic upstream selection produces a deterministic program. The program is then a software-daemon agent in its own right.
⟨ S, S ⟩ no persistent artifact
Sample per iteration
A runtime samples each next act from a learned distribution. Every iteration re-samples. No persistent artifact survives the iteration. Tool-use loops, reinforcement-learning policies at inference.
How agents move
Many actors switch cells across acts.
The actor axis says who. The selection axis says how. The substrate’s permitted-act set says what. Every agent known to the framework occupies one of these four cells at any given moment, and many actors switch cells across acts.
A human at a terminal alternates between ⟨H, D⟩ — typing a known command — and ⟨H, S⟩, deciding what command to type next. A language-model runtime alternates between ⟨S, S⟩ — sampling the next token — and ⟨S, D⟩, emitting a deterministic script when asked to. Movement across cells is the rule, not the exception.
The cells that produce durable artifacts — ⟨H, D⟩ and ⟨S, D⟩ — pay their selection cost once and then execute at the floor. The cells that sample per iteration — ⟨H, S⟩ and ⟨S, S⟩ — pay selection costs continuously. This asymmetry is the central economic claim of the framework.
Continue