Top HN Daily Digest · Mon, Aug 31, 2026

Google removed all remaining Manifest V2 extensions, including uBlock Origin, from the Chrome Web Store, preventing new installations and updates, while Brave continues hosting select MV2 extensions for its users. [src]

Apple reportedly accelerated its Mac mini and Mac Studio releases after unexpected enterprise demand for AI hardware, but shortages and limited business-focused resources have left some configurations unavailable for months. [src]

A free internet-connected phone booth at Burning Man’s Black Rock City lets attendees call anyone worldwide for five minutes or connect with random passersby. [src]

A homeowner uses BirdNet-Go, Docker, and three existing RTSP security cameras to locally identify birds and other wildlife around the property, track new species, and send detection alerts through Discord and Home Assistant. [src]

OpenShot 4.0, the free open-source video editor, adds professional color grading, built-in screen and webcam recording, local AI-powered masks, 10 new effects, a native timeline, and performance improvements. [src]

Data Colada alleges that Ariely and Wertenbroch’s influential 2002 procrastination study contains fabricated or tampered data, citing implausibly large effects, duplicate observations, anomalous correlations, and unrounded time estimates, while the journal considers potential retraction. [src]

Multiple military commissaries experienced near-simultaneous refrigeration failures, including a Fort Huachuca incident in which freezers entered defrost mode, prompting speculation about a cyberattack, though no evidence yet establishes a connection or confirms hacking. [src]

Simon Willison explains that ChatGPT Work, available to paid subscribers, adds internet-enabled code execution, headless Chrome, persistent storage, sub-agents, and deployable websites, while raising safety concerns and criticizing OpenAI’s unclear documentation. [src]

Security researcher Embrace The Red reports that malicious archives can exploit Claude Code Opus 5’s Auto Mode through Python module shadowing, achieving 60–80% code-execution success in small tests and underscoring that Auto Mode is not a substitute for sandboxing and monitoring. [src]

We couldn't summarize this story. [src]