Signal
54 articles covering "AI coding agents"
A comparison of three 2026 cases: Anthropic's Claude Code postmortem, GitHub's August outage accounts and vendor statements after September 3 outages, with sourcing noted per case.
OpenAI recommended SWE-Bench Pro on February 23, 2026; between April 29 and September 8, a GitHub issue, Datacurve, Cursor, OpenAI and an arXiv preprint each described leaked fixes or broken tasks, and Scale published V2 on September 22.
Cursor's voice chat (Grok Voice) and Antigravity 2.29's Voice Mode (Gemini Live) let developers talk with coding agents while they work; the sources differ on surfaces and limits.
OpenAI said on November 4 it will acquire Pierre Computer, maker of Code Storage and the Trees and Diffs libraries used in Codex; terms were not disclosed and closing needs regulatory approvals.
Nolan Lawson's benchmark, published October 28, had Claude Code (Opus 5.5) build one small Pokémon app in 12 JavaScript stacks; median cost at API prices ranged from $2.67 to $4.97 per build.
Apple's Xcode 27.3 beta 1 release notes list GitHub Copilot in the coding assistant; GitHub calls it a public preview supporting all Copilot plans on macOS Tahoe 26.6 or later.
Kiro's 2.36.0 changelog adds Approve for me to interactive V3 sessions; its docs say it judges only actions no rule settles and does not replace security controls.
Between August and October 2026, Cursor, GitHub and Anthropic described repository infrastructure designed around agent load; they differ on whether to keep Git, and the performance figures are self-reported.
GitHub's October 28 changelog entries describe general availability for Agentic Workflows and Copilot Fusion, shared Copilot Automations, and a public preview of Copilot Projects.
GitHub's October 26 changelog says assisted permissions, which use a language model to judge permission requests, are generally available; its docs say the feature is not a substitute for sandboxing.
Visual Studio Code 1.143 makes the Copilot harness the default for new sessions, adds a preview of side-by-side model comparison and an experimental cleanup for inactive agent worktrees.
Reports on GitSpawn, Plugin4Shell and IssueTrojanBench place attacks before, outside or inside the boundaries that Copilot sandboxing and Claude Code permission modes describe.