Memory layer bake-off on a 40-session support corpus
Agentic Memory System- proposed
- voting
- running
- measuring
- concluded
Preregistration · v2 · 25 Jun 2026 · immutable after start
Hypothesis
A structured episodic store with summarised recall beats raw vector recall on decision-consistency precision by at least 10 points at 50k+ tokens of history, at no worse latency.
Kill condition
Kill if episodic+summary trails vector recall on precision by day 14 on the 50k+ split, or if its p50 retrieval latency exceeds 1.5× vector recall at any history length.
Result
ValidatedValidated. Episodic+summary beat vector recall by 19 points precision at 50k+ tokens and ran at 0.6× its p50 latency. Graph memory won only on entity-lookup. Consolidation cut repeat-decision variance from 31% to 8%. Published as r-memory-layer; g-vector-memory-v1 superseded.
Running notes · fed from harness sessions and by the pair
- manual21 Aug 2026TO Tom Okafor
Concluded. Latency prediction was wrong in our favour: summarised recall is faster, not slower, because the index is smaller. Write-up to r-memory-layer.
- manual12 Aug 2026MT Mei Tanaka
Repeat-decision variance: 31% re-derived differently without consolidation, 8% with. That is the number Claire will care about.
- harness30 Jul 2026TO Tom Okafor
50k split: episodic 0.83, vector 0.64, vector+rerank 0.72, graph 0.69 precision. episodic p50 latency 0.6x vector. graph wins entity-lookup subset only
- harness14 Jul 2026TO Tom Okafor
adapter: 4 stores wired; corpus loaded; 220 questions; running 10k split
- manual25 Jun 2026TO Tom Okafor
Prereg v2: dropped the 'no-memory' baseline after the vote; it was not informative at 50k. Predicted outcome: episodic wins on precision, loses on latency, confidence 0.6.
Harness notes are auto-captured from Claude Code sessions: model, date, commit, session reference. Never the transcript, code or paths.