The numbers ledger

Every number, with its receipt

Every number I've cited, with the primary source, the date and how I checked it. ✓ means I read it at the source. Self-reported means a vendor said so and nobody audited it. Reuse any of it under CC BY 4.0. Just say where you got it.

Download: numbers.json · numbers.csv

NumberWhat it measuresSourceCheckedTrack · issue
>20%Extra inference cost from repository context files (AGENTS.md), with no general gain in task successGloaguen et al., ETH Zurich, arXiv:2602.11988
✓verified at sourceCKContext & knowledge
Issue #01
~66%Share of Cursor's system prompt trimmedCursor blog
self-reportedCKContext & knowledge
Issue #01
7%Cursor's total token reduction, "without degrading quality" (method not shown)Cursor blog
self-reportedCKContext & knowledge
Issue #01
22–25 of 27Top-four model scores on enterprise questions where the rules are statedEra by Eon, arXiv:2609.30055
✓verified at sourceCKContext & knowledge
Issue #01
≤6 of 24Scores of four of six models when the answer hinges on a hidden fact in contradictory recordsEra by Eon, arXiv:2609.30055
✓verified at sourceCKContext & knowledge
Issue #01
82%Estimated share of LLM migrations made after the model's shutdown date (22,555 commits, 17,703 GitHub repos)Kim, arXiv:2609.31288
✓verified at sourceUMUnderstanding & modernisation
Issue #01
94%LLM apps with the model ID hard-codedKim, arXiv:2609.31288
✓verified at sourceUMUnderstanding & modernisation
Issue #01
89% vs 13%Post-shutdown migrations under Anthropic's 60 to 114-day notices vs OpenAI's one-year noticeKim, arXiv:2609.31288
✓verified at sourceUMUnderstanding & modernisation
Issue #01
8%LLM migrations that switched model providerKim, arXiv:2609.31288
✓verified at sourceUMUnderstanding & modernisation
Issue #01
94.98%Best strict governance-gate success by an LLM (Gemini 3.8 Flash; GPT-5.6 Luna 83.29%, DeepSeek v4.1 Flash 74.18%; 300 synthetic projects, 899 runs)Canale, arXiv:2609.29345
✓verified at sourcePOPeople & operating model
Issue #01
1,700 of 1,700Governance gates passed by a deterministic rules baselineCanale, arXiv:2609.29345
✓verified at sourcePOPeople & operating model
Issue #01
0.5%Lost-ack writes duplicated by instructed frontier-model agents when a read-back was availableLi, LIMBO, arXiv:2609.29095
✓verified at sourceABAgents that build
Issue #01
56–74%Lost-ack writes duplicated by frontier-model agents with no read-backLi, LIMBO, arXiv:2609.29095
✓verified at sourceABAgents that build
Issue #01
28% → 4%Agent duplicate writes without vs with an idempotency key on every writeLi, LIMBO, arXiv:2609.29095
✓verified at sourceABAgents that build
Issue #01
40 of 111Agent approval records where the fields a human sees left effects uncoveredAgent Approval Laundering, arXiv:2609.28586
✓verified at sourceVTVerification & trust
Issue #01
1,000 PRs in 7 daysPRs merged by Kiro's team with 3 engineers and 50+ concurrent agent sessions (no quality metrics disclosed)Kiro blog
self-reportedABAgents that build
Issue #01

Last updated . Reuse with credit under CC BY 4.0: “The Dabbawala Protocol, thedabbawalaprotocol.com”.