{"data":{"event":{"id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","slug":"github-copilot-general-availability","title":"GitHub Copilot becomes a paid pair-programmer for every developer","short_summary":"On 21 June 2022 GitHub made Copilot generally available, the first mass-market AI coding assistant, powered by OpenAI Codex. A Microsoft/GitHub controlled study found developers completed a JavaScript HTTP-server task 55.8% faster with Copilot.","full_description":"Copilot had been a technical preview with more than a million developers. General availability sold the product at $10/month and, GitHub said, was already writing nearly 40% of code in files where it was enabled for popular languages. The underlying model was OpenAI Codex, a GPT-3 descendant fine-tuned on public code. Peng, Kalliamvakou, Cihon, and Demirer later reported a randomized experiment (n=95) in which treated developers finished a standardized task 55.8% faster (95% CI 21–89%). That study measures a narrow task, not firm-level output, and later field studies of AI coding tools have been mixed. The event is still the product launch that made inline AI coding a default part of professional software engineering, not a benchmark score and not a user-count trophy.\n\nUpdate (2026-09-06): Research acceleration: The view inside OpenAI ↗. Inside OpenAI, coding agents are reshaping AI research. Explore early data on agent usage, experiment velocity, task complexity, and research acceleration.\n\nUpdate (2026-09-21): AI coding has made CI a bottleneck, so we reworked ours to keep up. Article URL: https://linear.app/now/ci-bottleneck-reworked\nComments URL: https://news.ycombinator.com/item?id=49792067\nPoints: 14\n# Comments: 4","ledger_type":"benefit","primary_domain_id":"056250cf-05f4-4116-957c-78323ed7f8cf","event_status":"verified","event_date":"2022-06-21T00:00:00.000Z","discovery_date":"2022-06-21T00:00:00.000Z","first_published_date":"2022-06-21T00:00:00.000Z","last_reviewed_date":"2026-09-21T00:00:00.000Z","geographic_scope":"International","affected_population":"Professional software developers using Copilot","base_impact_tier":3,"base_score":"10.00","attribution_multiplier":"0.7500","evidence_multiplier":"0.5500","realization_multiplier":"0.8500","durability_multiplier":"0.7000","current_event_score":"2.454400","confidence_level":"high","score_explanation":"The first mass-deployed AI pair programmer is a major industry change (tier 3), not a new algorithm. Codex is the model inside the product (0.75). Evidence is a vendor-adjacent RCT plus GitHub telemetry, not independent firm-level productivity (0.55). Realization is a paid GA product (0.85). Follow-up ingest: added a follow-up source. Follow-up ingest: added a follow-up source.","methodology_version_id":"d7881163-fb23-4e71-8625-bac1a8662c0f","original_methodology_version_id":"d7881163-fb23-4e71-8625-bac1a8662c0f","published_at":"2022-06-21T00:00:00.000Z","created_at":"2026-09-13T21:49:40.117Z","updated_at":"2026-09-21T20:01:00.926Z","flags":[],"domain_name":"Computer Science","domain_slug":"computer-science","methodology_version":"0.1"},"contributions":[{"id":"d0417e5e-c168-434f-8ee1-cddb1b822685","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","model_id":"d96eca85-24db-4c1d-9a28-ca66dc5a2847","role_description":"GPT-3 descendant fine-tuned on public code; the model GitHub shipped as Copilot at general availability.","attribution_multiplier":"0.7500","credit_share":"1.0000","contribution_score":"2.454400","attribution_rationale":"GitHub Copilot GA was powered by OpenAI Codex; GitHub supplied the editor product and distribution.","attribution_confidence":"medium","first_used_date":"2021-06-29T00:00:00.000Z","model_version_if_known":null,"review_status":"approved","created_at":"2026-09-13T21:49:40.117Z","model_slug":"openai-codex","model_name":"OpenAI Codex","identity_class":"named_release","is_internal":false,"family_name":"Codex","family_slug":"openai-codex","organization_name":"OpenAI","organization_slug":"openai"}],"sources":[{"id":"59dc8b72-f570-4638-8756-7587414838b3","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","url":"https://arxiv.org/abs/2107.03374","canonical_url":"https://arxiv.org/abs/2107.03374","source_type":"preprint","publisher":"arXiv","author":null,"publication_date":"2021-07-07T00:00:00.000Z","retrieved_at":"2026-09-13T21:49:40.117Z","title":"Evaluating Large Language Models Trained on Code","excerpt":null,"content_hash":null,"source_reliability_class":"low","is_primary_source":true,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-13T21:49:40.117Z"},{"id":"37078e28-f1b6-4dae-9e53-acffc1c8df0a","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","url":"https://github.blog/news-insights/product-news/github-copilot-is-generally-available-to-all-developers/","canonical_url":"https://github.blog/news-insights/product-news/github-copilot-is-generally-available-to-all-developers/","source_type":"developer_announcement","publisher":"GitHub","author":null,"publication_date":"2022-06-21T00:00:00.000Z","retrieved_at":"2026-09-13T21:49:40.117Z","title":"GitHub Copilot is generally available to all developers","excerpt":null,"content_hash":null,"source_reliability_class":"low","is_primary_source":true,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-13T21:49:40.117Z"},{"id":"f33b5eb8-371f-45e4-8b1c-ced893d771c4","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","url":"https://arxiv.org/abs/2302.06590","canonical_url":"https://arxiv.org/abs/2302.06590","source_type":"preprint","publisher":"arXiv / Microsoft Research","author":null,"publication_date":"2023-02-13T00:00:00.000Z","retrieved_at":"2026-09-13T21:49:40.117Z","title":"The Impact of AI on Developer Productivity: Evidence from GitHub Copilot","excerpt":null,"content_hash":null,"source_reliability_class":"low","is_primary_source":true,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-13T21:49:40.117Z"},{"id":"b5fdb2c0-bee4-4b31-8c9a-924f39a78ddc","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","url":"https://openai.com/index/research-acceleration-view-inside-openai/","canonical_url":"https://openai.com/index/research-acceleration-view-inside-openai/","source_type":"developer_announcement","publisher":"OpenAI Alignment","author":null,"publication_date":"2026-09-06T00:00:00.000Z","retrieved_at":"2026-09-17T03:01:09.307Z","title":"Research acceleration: The view inside OpenAI ↗","excerpt":"Inside OpenAI, coding agents are reshaping AI research. Explore early data on agent usage, experiment velocity, task complexity, and research acceleration.","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-17T03:01:09.307Z"},{"id":"0dee5454-418a-48bc-a752-3e8b6c574adb","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","url":"https://linear.app/now/ci-bottleneck-reworked","canonical_url":"https://linear.app/now/ci-bottleneck-reworked","source_type":"journalism","publisher":"Hacker News","author":null,"publication_date":"2026-09-21T00:00:00.000Z","retrieved_at":"2026-09-21T20:01:00.911Z","title":"AI coding has made CI a bottleneck, so we reworked ours to keep up","excerpt":"Article URL: https://linear.app/now/ci-bottleneck-reworked\nComments URL: https://news.ycombinator.com/item?id=49792067\nPoints: 14\n# Comments: 4","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-21T20:01:00.911Z"}],"claims":[{"id":"afa31793-96e9-4711-be3e-65819d314806","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","claim_text":"GitHub made Copilot generally available to individual developers on 21 June 2022, powered by OpenAI Codex.","claim_type":"outcome","claim_status":"supported","confidence_score":null,"created_at":"2026-09-13T21:49:40.117Z","updated_at":"2026-09-13T21:49:40.117Z"},{"id":"4d694fc5-9de8-4d57-ba00-db74d1c68b40","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","claim_text":"In a 95-developer controlled experiment, Copilot users completed a JavaScript HTTP-server task 55.8% faster than the control group.","claim_type":"outcome","claim_status":"supported","confidence_score":null,"created_at":"2026-09-13T21:49:40.117Z","updated_at":"2026-09-13T21:49:40.117Z"},{"id":"809d968e-ee2f-4a5f-81f8-353991e4fa8e","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","claim_text":"The 55.8% speedup is a measured increase in firm-level software output.","claim_type":"significance","claim_status":"disputed","confidence_score":null,"created_at":"2026-09-13T21:49:40.117Z","updated_at":"2026-09-13T21:49:40.117Z"}],"revisions":[{"id":"b2482ac5-0059-45f1-9050-3b61e1ccd355","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","model_id":null,"previous_score":"0.000000","new_score":"2.454400","previous_factors":{},"new_factors":{"evidence":0.55,"base_score":10,"durability":0.7,"attribution":0.75,"realization":0.85},"change_reason":"Imported September 2026 mathematics events under methodology 0.1 credit-unit rules.","trigger_type":"editorial_import","trigger_source_ids":null,"reviewer_id":null,"review_status":"published","created_at":"2026-09-13T21:49:40.117Z"},{"id":"34b1a622-d949-4fdb-82b6-5018a003913c","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","model_id":null,"previous_score":"2.454400","new_score":"2.454400","previous_factors":{"tier":3,"evidence":0.55,"base_score":10,"durability":0.7,"attribution":0.75,"realization":0.85},"new_factors":{"tier":3,"models":["d96eca85-24db-4c1d-9a28-ca66dc5a2847"],"evidence":0.55,"base_score":10,"durability":0.7,"attribution":0.75,"realization":0.85},"change_reason":"Ingest update: added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["b5fdb2c0-bee4-4b31-8c9a-924f39a78ddc"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-17T03:01:09.321Z"},{"id":"6fb362f2-539f-450b-8a0f-9c76fa118c1d","event_id":"f3b570fe-bbd9-4632-bf1c-c57d6f4aeee0","model_id":null,"previous_score":"2.454400","new_score":"2.454400","previous_factors":{"tier":3,"evidence":0.55,"base_score":10,"durability":0.7,"attribution":0.75,"realization":0.85},"new_factors":{"tier":3,"models":["d96eca85-24db-4c1d-9a28-ca66dc5a2847"],"evidence":0.55,"base_score":10,"durability":0.7,"attribution":0.75,"realization":0.85},"change_reason":"Ingest update: added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["0dee5454-418a-48bc-a752-3e8b6c574adb"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-21T20:01:00.949Z"}],"secondary":[{"name":"Engineering","slug":"engineering"}]},"methodology_version":"0.1"}