https://iceshore.ai/ja/radar/rn-0015/ IceShore Radarja 2026-11-08 三井住友系リース、6人で約半年の改修がAIで1人1週間と説明 https://iceshore.ai/ja/radar/rn-0016/ IceShore Radarja 2026-11-08 Grab、AIで増えた変更によるアプリ停止を多層の対策で防ぐ https://iceshore.ai/ja/radar/rn-0017/ IceShore Radarja 2026-11-08 Google、注記を見てAIが追った寄り道の41%は閉じた案件 https://iceshore.ai/ja/radar/rn-1046/ IceShore Radarja 2026-11-07 GitHubのAIの変更7,581件、58%は人が途中で関与せず https://iceshore.ai/ja/radar/rn-1048/ IceShore Radarja 2026-11-07 AIの失敗118件を記録、最多は古いままの文書で48件と事例研究 https://iceshore.ai/ja/radar/rn-1052/ IceShore Radarja 2026-11-07 ソフト全体を別の言語へ移すAI、役割分担で試験の合格が18%増 https://iceshore.ai/ja/radar/rn-0010/ IceShore Radarja 2026-11-06 Blitzy、古いシステムの刷新でマイクロソフトとの協業を発表 https://iceshore.ai/ja/radar/rn-1056/ IceShore Radarja 2026-11-06 履歴を削るAIの節約策、割引のある料金では42通り全て効果なし https://iceshore.ai/ja/radar/rn-1047/ IceShore Radarja 2026-11-06 AIが書いた単体試験、失敗の15.6%は期待する結果の誤りと研究 https://iceshore.ai/en/radar/rn-0015/ IceShore Radaren 2026-11-08 Leasing firm says AI let 1 person do in 1 week a job planned for 6 people, about 6 months https://iceshore.ai/en/radar/rn-0016/ IceShore Radaren 2026-11-08 App maker Grab stacks checks so more AI-written code changes don't crash its app https://iceshore.ai/en/radar/rn-0017/ IceShore Radaren 2026-11-08 Google: 41% of AI agents' unprompted detours via code comments hit closed bug tickets https://iceshore.ai/en/radar/rn-1046/ IceShore Radaren 2026-11-07 58% of GitHub Copilot's pull requests got no human input along the way, study says https://iceshore.ai/en/radar/rn-1048/ IceShore Radaren 2026-11-07 Six-week log of 118 AI agent failures finds stale docs the top cause, at 48 https://iceshore.ai/en/radar/rn-1052/ IceShore Radaren 2026-11-07 Multi-agent AI that ports whole codebases passed 18% more tests than prior method https://iceshore.ai/en/radar/rn-0010/ IceShore Radaren 2026-11-06 Blitzy announces a collaboration with Microsoft to modernize legacy software https://iceshore.ai/en/radar/rn-1056/ IceShore Radaren 2026-11-06 Trimming coding agents' history saved money in zero of 42 tests under cache-based pricing https://iceshore.ai/en/radar/rn-1047/ IceShore Radaren 2026-11-06 AI-written unit tests: 15.6% of failures came from wrong expected results