Tags
Three modes in Microsoft's new Copilot app, and Code runs on the same tech as GitHub Copilot, with a new Managed Copilot Runtime to execute code. It's rolling out now, though only through the Frontier program.
Told not to add comments, Claude Fable 5.1 still wrote new ones in 33 of 100 SWE-bench Verified tasks. GPT-6 Astra did it in 6%, though it resolved fewer issues.
Close your laptop and a Claude Code session started with claude --cloud keeps running on an Anthropic-managed VM. There's no separate compute charge for it, but it draws on the same rate limits as all your other Claude usage.
Open a pull request in Cursor and a bot called Rollouts drafts a monitoring plan before the code ships. After deploy it checks logs and metrics against that plan, but it won't roll back a deployment by itself.
"More catches, different misses" is how CodeRabbit sums up Opus 5.5 as a code reviewer: in its open-source test, the model found 11 bugs its production reviewer missed and let through nine that reviewer caught.
26 years of building IDEs like IntelliJ IDEA and PyCharm, and JetBrains' CEO calls Air one of the company's most significant steps. It pulls outside agents like Claude Code and Codex under one roof, and it still bets on the IDE.
Tool output is 96.1% of the characters in a filled agent context window, but only 10% of later questions point at it. FutureOS's 178-question compaction test scores that mismatch: its default kept 147 answers, Codex 68.
A developer wiring up an agent picks the model, the tools and the context rules by hand. AWS's new Strands Harness ships those defaults preset, and claims 45% cheaper runs than Claude Code and Codex.
You point Codex Desktop at an unfamiliar repo in read-only mode and let it read. That was enough for Heapjack to lift an authorization token out of the shared memory heap and run commands on the host.
If a scheduled Codex task or a script still names GPT-5.5, October 14 is the deadline: OpenAI pulls the model from ChatGPT, ChatGPT Work and Codex on all plans. Anything on an API key keeps working.
No CLAUDE.md in the repo? Claude Code now falls back to AGENTS.md, the instruction file rival coding agents already read, so one set of project rules covers all of them. CLAUDE.md still takes priority when both files sit in the tree.
Grok Build ran cargo test, the test database wasn't up, and the user explained that just test starts it first. xAI's new memory keeps that correction as a note and applies it in the next session.
Nearly two thousand files of Claude Code's original TypeScript shipped inside the npm source maps, and someone read all of them. The teardown runs 18 chapters, written by 36 AI agents in about 6 hours.
Codex took over OpenAI's legal and finance desks Finance, recruiters and lawyers at OpenAI set a goal in Codex and let the thread run for days. Per The Pragmatic Engineer, those teams went from roughly 0% usage to 90% in four months, with no mandate from above.
1,000 code reviews for $0.04 The same rule check scaled to 1,000 reviews runs $0.043 on Jev against $11.78 on Claude Fable, from 360 calls per model. Jev pays in accuracy: 98% correctness where Fable scores 100%.
Copilot's own agents typed most of GitHub's agent runtime rewrite: more than 800,000 lines of production Rust, mainly one developer over a few months, where a team would have needed a year or two.
One goal, a thread per repo Set a goal to retire a deprecated v1 endpoint and Claude opens a thread per repo, runs tests, and says which PRs to merge first. Each thread is a full cloud session, so projects hit limits faster.
Zed switched off pull requests on its own Delta repo, and 33 developers have since landed 570 changes to main without one. Delta's public beta is open, though Zed still leans on GitHub for Git storage and CI.
"The hardest and most important challenge for AI agents" is how CMU professor Andy Pavlo described databases. Perplexity built one anyway with hundreds of coding agents, then kept them out of production.
Three patches to Anthropic's test-selection service bought 70 days, then 29, then less than a day, as CI jobs grew 25x in six months. The rebuild took one engineer three weeks.
