Thursday, May 7, 2026
AI beats doctors in the ER (seriously)
Apple's planning to let iOS 27 users pick their AI model (bold move), while image AI apps are crushing text-only ones with 6.5x more downloads. Meanwhile, a Harvard study shows AI actually beat two human doctors at ER diagnoses—though that raises some wild questions about where this is all heading. Would you trust AI over your doctor in the ER?
Top Stories
TechCrunch
Image generation models are driving 6.5x more AI app downloads than text models, with Gemini and ChatGPT adding tens of millions of installs after image model releases, though only ChatGPT successfully converted downloads into significant revenue ($70M in 28 days).
TechCrunch
Apple is developing a feature for iOS 27 that will let users choose between multiple third-party AI models, including those from Google and Anthropic, to power Siri and other on-device functions. This reflects Apple's hardware-focused AI strategy as new CEO John Ternus takes the helm.
9to5Mac
OpenAI released ChatGPT for Intune, a separate iOS app designed for enterprise and school users requiring Microsoft Intune device management, enabling organizations to deploy ChatGPT while maintaining IT security and compliance standards.
OpenAI's o1 model outperformed internal medicine physicians in diagnosing 76 real emergency room cases, achieving 67% accuracy versus 50-55% for human doctors, though researchers stress the need for real-world trials and note limitations including comparison against non-ER specialists.
Google Blog
Google's Gemini API File Search now supports multimodal RAG with image and text processing, custom metadata filtering, and page-level citations for improved accuracy and verifiability in AI applications.
Keep Reading
Industry Voices
Enjoyed this issue?
Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.