Loading the edition…
· Past edition · 16 stories
Retrospective edition reconstructed from later captures. Paper versions may include later updates; coverage labels describe the evidence used.
Applications & products
Claude Cowork’s desktop app now has a separate browser for delegated website tasks, while Claude in Chrome remains suited to pages already open in your browser.
Why it matters: Choosing the browser based on the task can keep your existing browser separate while Claude handles standalone web work. But safeguards do not eliminate prompt-injection risk, so the article recommends starting on trusted sites.
Applications & products
Open each story in the reader. Summaries are shown when available; some sources provide an abstract only. With an account you can keep your own notes.
OpsHarness argues that root-cause analysis should improve by evolving the harness around a capable general agent, rather than rebuilding the agent—and reports 59.0% top-1 accuracy.
Why it matters: For teams deciding where to invest, the paper's argument is to reuse a capable general agent and improve the system-specific harness. Updates pass through dual-gate verification designed to prevent overfitting and regression; the abstract does not establish that these problems are guaranteed to be prevented.