Two-timescale control loops

Separates fast inference or parameter updates from slower structural adaptation, policy review, pruning, or mode selection.

Research
Last verified
Not verified
Updated
Reading time
1 minutes

Separates fast inference or parameter updates from slower structural adaptation, policy review, pruning, or mode selection.

Fast loop

Examples include token inference, classification, score updates, or tightly bounded parameter changes. The fast loop should have a predictable per-step budget.

Slow loop

Examples include accepting a new rule, resizing a structure, pruning state, changing model or precision, compacting memory, or entering a halted mode.

Coordination

Record which state is shared, when transitions are permitted, how concurrent work is stopped, and which invariants must hold before and after a structural change.

Scope

This starter page defines the questions, boundaries, evidence, and failure modes that should be recorded before a capability is presented as supported.

Engineering considerations

  • Identify the source, version, target environment, and owner.
  • Separate observed values from estimates and externally reported values.
  • Record trade-offs, unsupported cases, and fallback behavior.
  • Link performance statements to a compatible benchmark methodology.

Verification questions

  • What exact artifact, revision, backend, and environment were reviewed?
  • Which assumptions could change the result?
  • Which data should be retained so another engineer can reproduce the conclusion?