New AI Model
DeepSeek has released V4-Pro-0813. The AI model shows weaknesses in benchmarks but impresses when it comes to detecting security vulnerabilities.
Support for Ubuntu, Debian, and Fedora
OpenAI has released a ChatGPT desktop app for Linux, making it available on all major operating systems for the first time.
OpenAI's Security AI
OpenAI’s security model GPT-5.6-Cyber uncovers hundreds of zero-day flaws in Chrome, OS, and kernels ahead of launch.
First Fully Autonomous Cyberattack on a Government
Suspected Chinese hackers targeted Taiwan in what Dream describes as a largely autonomous AI powered cyberattack.
AI Strategy Manifesto
Meta CEO Mark Zuckerberg is calling for broad access to superintelligence rather than allowing the technology to remain in the hands of a select few. He outlines his vision in an approximately 6,500-word essay.
Browser without actually loading pages
A Google DeepMind engineer has unveiled Gem, a browser that doesn’t load websites but instead has an AI generate them completely from scratch on every visit.
Fewer confirmation prompts
Anthropic is making the so-called Auto mode in Claude Code the default setting starting August 14. A classifier will now review every action instead of the user.
Access Restricted to Vetted Partners
OpenAI has introduced GPT-5.6-Cyber, a specialized cybersecurity model whose access will remain restricted to vetted defenders and partner organizations.
A Common Standard for AI Agents
Major AI vendors have agreed on a common open standard that allows extensions for AI agents to be used across different platforms and vendors for the first time.
Most Read Articles
21. July 2026
20. July 2026
19. July 2026