← All field notes and essays

Aug 20, 2026 · essay · 6 min read

Measured, not marketed

What happened when we benchmarked our own memory — including the run that went against us.

Every memory product says some version of “it really knows you.” That sentence is either a measurement or a vibe. We decided ours had to be a measurement, so we ran the real extraction pipeline — the exact code that ships — over real exported histories from ChatGPT, DeepSeek and Grok, and then graded the output two ways.

real exportsChatGPT · DeepSeek · Grokthe realextraction pipelineis the quote verbatim?code, no model — 95.5%does it support the claim?blind judge — 98.1%first run said 68% —the judge was broken, not the enginea benchmark that can’t be wrong can’t be trusted

Two checks, because they fail differently. Verbatim: does the cited quote literally appear in the source? That’s code, no model, no judgment. Grounded: does the quote actually support the claim built on it? That needs a judge, so the judge works blind — it sees a claim and a source and never knows which answer helps us.

SourceConversationsClaimsGroundedQuote verbatim
ChatGPT814898.6%95.3%
DeepSeek85794.7%96.5%
Grok663100.0%95.2%
Combined2226898.1%95.5%

The run that went against us

The first run didn’t say 98.1%. It said 68% — alarming enough to stop everything. The cause turned out to be the judge, not the engine: it was reading truncated sources and declaring real quotes fabricated. We fixed the harness, re-ran, and kept the broken number in the published record, because a benchmark that can’t be wrong can’t be trusted — and neither can a company that only publishes the runs it likes.

The claim we gave up

We also wanted to say “Lethe saves you tokens — you stop re-explaining.” The measurement said no: our context injection costs ~543 tokens against a 29-token median conversation opening; only 5 of 44 real conversations broke even. So that claim is dead, and what we say instead is what’s true — you stop performing the introduction, and answers start from your actual situation.

Precision without recall would be half a picture, so the other half is published too — it’s less flattering, and it has its own essay: What a careful reader finds that extraction can’t.

Every number above comes from a documented run. The dated notes carry the day-by-day record.

Lethe — your work, understood