AI
OpenAI launches ChatGPT Health and brings Voice to the desktop app
OpenAI rolled out two significant ChatGPT expansions on Thursday. ChatGPT Health is now reaching U.S. users, letting them securely connect Apple Health data and supported medical records so the model can contextualize results, track changes over time, and hold more informed conversations. OpenAI said more than 300 million people already bring health-related questions to ChatGPT each week, and the feature was built with input from hundreds of physicians who helped measure accuracy, safety, and communication.
Separately, ChatGPT Voice is now available in the desktop app globally, powered by GPT-Live. Users can speak to control their computer and direct multiple agents running inside ChatGPT Work or Codex simultaneously, with the model able to listen, speak, and coordinate tasks in parallel. OpenAI developers also confirmed that Sites Analytics has entered public testing, giving Codex-published websites basic performance metrics.
Google DeepMind unveils Gemini 3.5 Flash Cyber, a security-specialized lightweight model
Google DeepMind announced Gemini 3.5 Flash Cyber, a lightweight model purpose-built for security teams to detect and patch vulnerabilities before they are exploited. The model is designed for speed and efficiency, allowing analysts to evaluate significantly more code paths than with general-purpose models. DeepMind said internal testing on its own codebases, including Chrome and Android, showed the model consistently caught complex, novel vulnerabilities that standard models missed. Given its sensitivity, the rollout is starting with a limited-access pilot for governments and trusted partners.
This briefing is part of the archive.
Get the app for full archive access — every daily briefing, sourced and attributed.