This Thursday morning, multiple reports confirm a significant shift in AI capabilities, particularly in cybersecurity and healthcare.

Both GeekWire and streamlinefeed.co.ke highlight that frontier AI models are now matching human cybersecurity professionals. Microsoft's new AI system, MDASH, scored 88.45% on the CyberGym benchmark, outperforming Anthropic’s Mythos Preview at 83.1% and GPT-5.5 at 81.8%. MDASH has already found 16 new vulnerabilities in Windows. The UK AI Security Institute confi