Release Report

What a memory system actually does, what it costs, and what it still gets wrong — assembled from measurements rather than summarised.

Why it matters in a memory layer

A retrieval system ships against a benchmark: a score on a shared corpus, comparable to somebody else's. It is a weak claim and it is legible, and there is industry-wide agreement about what it means.

There is no such agreement here — memory benchmarks share neither their corpora nor their division of labour with the reading model. So the honest artifact is not a number. It is the measurements plus the gaps, each naming the lesson that found it, so a reader can reproduce any claim and see the shape of what was not attempted.

Three exams answer different questions and all three are needed: does the store believe the right thing, would it ever say it, and does it survive a tight context. A system can pass the first and fail the second, which is a test in this course's suite for exactly that reason.

Connections

Taught in: Hardening Pass

Do not confuse with: Benchmark Claim