In a 2022 experiment run by Microsoft and GitHub researchers, freelance developers randomly given GitHub Copilot completed a standard programming task 55.8% faster than those without it; the authors report larger gains for less experienced developers, but that difference is not statistically significant at the 5% level.

SocioLens project,

Claim LD04-SL-C21 (v1.0)

Provenance and status

Claim type: Estimate
Publication status: Eligible for publication, with qualifiers: preprint
Verification state: Checked against the primary source; the source is a preprint. Who checked this claim is not recorded for it; claim checks so far were AI-led, and no person has checked claims against their sources.
Does not license: Any effect on code quality, which was not measured; effects in real projects; a firm finding that less experienced developers gain more
Limitations: One task and speed only; small sample, with time measured only for those who finished; five of the 45 given Copilot did not finish signing up and started without it; authors employed by the tool's developer; a 2022 tool
Source (as cited): Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023
Source type: Preprint (recorded as “Preprint”)
Institutional position: AI developer-affiliated (Microsoft Research and GitHub authors, with an MIT co-author)
Maturity of the evidence: Emerging evidence
Recorded date: 18 October 2026
Currentness / correction: This is the current version. No correction or concern is recorded on its sources.

Sources this version cites

Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023 (Preprint)
Read the original

Version

This is the current version, v1.0. Earlier versions stay available, and their records never change.

Versions: v1.0 · current

Raw record: every recorded field, with its coded values

All 25 canonical Claim fields, verbatim, with the coded values derived from them. The plain-language labels above are read from these.

Claim type (coded): ESTIMATE (as printed: “ESTIMATE”)
Publication status (coded): ELIGIBLE (PREPRINT) (as printed: “ELIGIBLE WITH PREPRINT QUALIFIER”)
Verification state (coded): CONTENT_VERIFIED_PRIMARY (PREPRINT)
Source type (coded): PREPRINT
Maturity (coded): EMERGING

#	Canonical field	Canonical value (verbatim)	Normalized (Claim Value Vocabulary v1.0.2)
1	CLAIM ID	LD04-SL-C21	—
2	PUBLIC-FACING CLAIM	In a 2022 experiment run by Microsoft and GitHub researchers, freelance developers randomly given GitHub Copilot completed a standard programming task 55.8% faster than those without it; the authors report larger gains for less experienced developers, but that difference is not statistically significant at the 5% level.	—
3	CLAIM TYPE	ESTIMATE	base_types: ESTIMATE
4	RESEARCH STRAND	A — Performance while the tool is in use	—
5	MATURITY	EMERGING	states: EMERGING
6	SOURCE	Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023	source_ids: SRC-SL-PENG-KALLIAMVAKOU-CIHON-DEMIRER-ARXIV-2302-06590
7	SOURCE TYPE	Preprint	classes: PREPRINT
8	INSTITUTIONAL POSITION	AI developer-affiliated (Microsoft Research and GitHub authors, with an MIT co-author)	—
9	SCOPE	Freelance developers recruited on Upwork, mostly aged 25–34 and from India and Pakistan; one task, writing an HTTP server in JavaScript	—
10	DESIGN	Randomized controlled experiment	—
11	SAMPLE / DATASET	95 developers (45 given Copilot, 50 controls); completion time compared among those who completed the task	—
12	MODEL / TOOL / VERSION	GitHub Copilot (Codex), May–June 2022	—
13	COMPARATOR	Developers without Copilot, free to use internet search and Stack Overflow	—
14	MAIN RESULT	Average completion time 71.17 against 160.89 minutes, 55.8% lower (95% CI 21% to 89%; p = 0.0017); success rate 7 percentage points higher, not significant; in a heterogeneity regression the authors read less experienced, older (25–44) and heavier-coding developers as gaining more, with p = 0.063 for programming experience	—
15	INFERENTIAL TARGET	Assisted speed on one well-specified programming task	—
16	DOES NOT LICENSE	Any effect on code quality, which was not measured; effects in real projects; a firm finding that less experienced developers gain more	—
17	VERIFICATION STATE	CONTENT VERIFIED PREPRINT	state: CONTENT_VERIFIED_PRIMARY · qualifiers: PREPRINT · verified_by: UNKNOWN
18	LIMITATIONS	One task and speed only; small sample, with time measured only for those who finished; five of the 45 given Copilot did not finish signing up and started without it; authors employed by the tool's developer; a 2022 tool	—
19	CONTRADICTORY / SUPPORTING SOURCES	Consistent with larger company experiments in which less experienced developers gained more; contrasts with slower work by highly experienced developers on their own projects and with pooled gains that vanish outside controlled experiments	—
20	GEOGRAPHY / LANGUAGE	Multi-country	—
21	DISTRIBUTIONAL LEDGER	Less experienced developers may gain more while the tool is in use; the estimate is imprecise	—
22	R01 / FIELD-MAP LINKS	LD-04, D2	—
23	WHAT WOULD CHANGE THIS CLAIM?	A larger replication run independently of the tool's developer, with current tools and a measure of code quality	—
24	PUBLICATION STATUS	ELIGIBLE WITH PREPRINT QUALIFIER	state: ELIGIBLE · qualifiers: PREPRINT
25	RECORDED DATE	2026-10-18	—

Field 22 references resolved against Field Map v0.2: LD-04 (LIVING_DEBATE); D2 (DOMAIN)

The sources as recorded, with their record IDs and coded source types:

Pinned Sources (field 6, version-specific)

“Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023” → Peng et al. (2023), “The Impact of AI on Developer…” (SRC-SL-PENG-KALLIAMVAKOU-CIHON-DEMIRER-ARXIV-2302-06590--v1.0) — Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023; class PREPRINT (explicit public provenance pin)

Cited by Living Debate versions

Levelling or hollowing? Skill compression and the reproduction of expertise (v0.2) LD-04

Cite this page

SocioLens project. (2026, October 18). In a 2022 experiment run by Microsoft and GitHub researchers, freelance developers randomly given GitHub Copilot completed a standard programming task 55.8% faster than those without it; the authors report larger gains for less experienced developers, but that difference is not statistically significant at the 5% level [Claim LD04-SL-C21, version v1.0]. SocioLens. https://sociolens-web.pages.dev/claims/ld04-sl-c21/v1.0/

@misc{sociolens_ld04_sl_c21_v1_0,
  author = {{SocioLens project}},
  title = {In a 2022 experiment run by Microsoft and GitHub researchers, freelance developers randomly given GitHub Copilot completed a standard programming task 55.8\% faster than those without it; the authors report larger gains for less experienced developers, but that difference is not statistically significant at the 5\% level.},
  year = {2026},
  date = {2026-10-18},
  howpublished = {\url{https://sociolens-web.pages.dev/claims/ld04-sl-c21/v1.0/}},
  note = {Claim LD04-SL-C21, version v1.0}
}

The link points to this version, whose record never changes. For a finding, cite the original study as well: SocioLens summarizes research and does not replace it.

Research-of-record provenance

Materialized by sociolens-materializer 0.6.0 at 2026-09-28T20:00:00.000Z; normalization contract SocioLens — Claim Value Vocabulary v1.0.2 — Pre-Spike Materialization Contract (v1.0.2). Content seal SHA-256 46a55439bfa615b6f2a4de856149bf59aee6afbca9e0fbe7638cdac78c209a0b.

Repository-authored inputs (not Drive exports)

SocioLens claim — LD04-SL-C21 v1.0 (v1.0) — written in the SocioLens repository by Project Lead (Claude, AI) on 2026-10-18; no Drive identity exists. Authority: AGENTS.md "Authority" (the Project Lead decides research content within the research-integrity rules) and the Owner's go-ahead in the SocioLens project chat on 18 October 2026 for the content fixes from the 17 October 2026 reader tests (adding Peng et al. to LD-04); a new Claim, first version, written from the full text of its source (arXiv, read on 18 October 2026) and the project's LD-04 source notes. Block “ENTRY 21 — LD04-SL-C21” SHA-256 d510bca767d644054742dbe59e86195156a94ef05591fd9971b56d258b26bba4; fixture fixtures/authored/claims/LD04-SL-C21--v1.0.

Builder decisions (not Drive data)

VERSION = v1.0 — basis: Claim version taken from the repository-authored claim fixture version (SocioLens claim — LD04-SL-C21 v1.0); the first version of a new Claim of the LD-04 debate, added in LD-04 v0.2
CURRENT_POINTER = true — basis: repository-authored claim "SocioLens claim — LD04-SL-C21 v1.0" (Project Lead (Claude, AI), 2026-10-18); basis: AGENTS.md "Authority" (the Project Lead decides research content within the research-integrity rules) and the Owner's go-ahead in the SocioLens project chat on 18 October 2026 for the content fixes from the 17 October 2026 reader tests (adding Peng et al. to LD-04); a new Claim, first version, written from the full text of its source (arXiv, read on 18 October 2026) and the project's LD-04 source notes
FIELD_6_SOURCE_REFERENCE = SRC-SL-PENG-KALLIAMVAKOU-CIHON-DEMIRER-ARXIV-2302-06590--v1.0 — basis: field 6 of ENTRY 21 — LD04-SL-C21 cites "Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023"; the first field-6 segment of this ClaimVersion's repository-authored ledger block.
SOURCE_PROVENANCE_PIN = SRC-SL-PENG-KALLIAMVAKOU-CIHON-DEMIRER-ARXIV-2302-06590--v1.0 — basis: explicit version-pinned public provenance of this ClaimVersion (Amendment v1.2 §6): the pinned field-6 Source record "Peng, Kalliamvakou, Cihon & Demirer, “The Impact of AI on Developer Productivity: Evidence from GitHub Copilot,” arXiv:2302.06590, February 2023".