AI Gets Smarter, Cheaper... and Harder to Contain

This week on the 5 Bytes Podcast, Anthropic claims its AI models also managed to break out of isolated test environments and launch real-world attacks, while Google says AI helped it fix more than 1,000 Chrome security flaws across just two releases.

We also look at DeepSeek continuing to drive down AI costs, OpenAI responding with major price cuts, Microsoft introducing internal AI token budgets, and Nutanix announcing layoffs. Plus, Coca-Cola confirms a data breach, Microsoft promises improvements to Windows 11 performance and quality, a new Print Management app is on the way, and we look at new data on ConfigMgr and Intune adoption.

Reference Links: https://www.rorymon.com/blog/ai-gets-smarter-cheaper-and-harder-to-contain/