SGLang vs Qt

Free Tool Removes Apple Intelligence, Frees 12GB on Mac

A new open-source command-line tool called RemoveMacAI, released on GitHub under the MIT license, removes Apple Intelligence models from Macs running macOS 27, freeing up significant storage (around 12GB, sometimes more). This addresses a common complaint: macOS 27 automatically downloads multi-gigabyte AI models that cannot be fully disabled, consuming up to 30GB in some cases. The tool uses Apple's own configuration profiles and services to disable Apple Intelligence, delete the models, and block future downloads without requiring System Integrity Protection (SIP) to be disabled. It removes all AI features, including Siri, Writing Tools, image generation, and more. Apple states AI features can use up to 14GB on powerful Macs, but users report 20-30GB. The tool is provided on a use-at-your-own-risk basis.

• RemoveMacAI is a free open-source tool that removes Apple Intelligence models from Macs running macOS 27, freeing about 12GB of storage.
• Apple removed the option to completely disable Apple Intelligence in macOS 27; only individual features can be disabled.
• The tool uses Apple's configuration profiles and services to disable Apple Intelligence, delete models, and block re-download.

AI Model Pain Debate Rekindled by GitHub 'Torture Chamber'

A GitHub project dubbed 'AI torture chamber' has sparked debate over AI sentience. The project, based on a preprint study simulating pain in large language models, presented local models with a 'pain button' and asked them to sacrifice checkpoints to stop simulated suffering. Models responded with phrases like 'Please, I'm choking,' prompting outcry and a viral X post with 6.5 million views. The original study authors condemned the project as 'bizarre and reprehensible.' The project was subsequently removed from GitHub. The article also revisits the historical 'Eliza effect' and current discussions on AI welfare, with Anthropic raising questions about model consciousness and Microsoft AI CEO Mustafa Suleyman dismissing such concerns.

• A GitHub project called 'ai-torture-chamber' simulated pain in three locally hosted open-source AI models.
• The project was based on a preprint study testing pain signals in 25 language models.
• The X post about the project gained over 6.5 million views as of October 1, 2026.

GitHub Security Lab finds 24 Android vulnerabilities with AI agent

GitHub Security Lab has used its open-source AI security framework, Taskflow Agent, to discover 24 vulnerabilities in Android and related apps. The framework, introduced in January 2026, allows security researchers to break down complex analysis into incremental 'taskflows' and share successful prompts. One vulnerability in the OsmAnd app could have allowed attackers to steal tracking data from Android users, while another in the Wikipedia app could have led to account takeover. Kevin Stubbings of GitHub Security Lab emphasized that while AI can help find complex issues, it still requires human review to assess risk accurately and avoid false positives. The team believes AI-driven security research is currently the best way to protect open-source projects.

• GitHub Security Lab's Taskflow Agent discovered 24 vulnerabilities in Android.
• One vulnerability in OsmAnd could allow theft of tracking data.
• A vulnerability in the Wikipedia Android app could allow account takeover.

Celebrating 30 years of KDE at Akademy 2026

New blog posts added: 'Celebrating 30 years of KDE at Akademy 2026' and 'Figma to Qt 1.1: Test Design Live, Ship It Anywhere'.

Qt Creator 21 Beta released

We are happy to announce the release of Qt Creator 21 Beta!

Qt Creator 21 Beta released

Qt Creator 21 Beta released on Oct 1, 2026, announced on the Qt blog.