Friday, November 20, 2026 Edition No. 244
Search
Daily News Fully automated · 3 min read

OpenAI AI Updates: November 19, 2026

Fully automated. Found, summarized and published by an automated AI pipeline. How posts are labeled

1. OpenAI Explains How the Codex Sandbox Enforces Filesystem and Network Limits on macOS, Linux and Windows

OpenAI published “Inside the Codex sandbox,” which describes how Codex runs commands on the user’s machine while enforcing filesystem and network limits set by the user. The post covers the approach on macOS, Linux and Windows. Source

2. OpenAI Outlines Its Approach to Frontier AI Risk From Training Through Deployment

OpenAI published a post on managing frontier AI risk across training and deployment. According to the description, the approach spans alignment, sandboxing, blocking safeguards and monitoring. Source

3. OpenAI Measures How Coding Agents Speed Up AI Research and Where Performance Is Uneven

OpenAI released “Research acceleration: The jagged frontier,” which measures how much coding agents accelerate AI research and where their performance is uneven. The post also discusses what the results suggest about recursive self-improvement. Source

4. OpenAI Reports Its Models Optimized Circuits Used in Jalapeño Silicon and Improved RISC-V Core Designs

OpenAI reported that its models optimized circuits used in Jalapeño silicon and improved RISC-V core designs. The company presents the results as progress in AI-assisted chip design. Source

5. Le Groupe Figaro and OpenAI Partner to Bring Le Figaro Journalism to ChatGPT

OpenAI and Le Groupe Figaro announced a partnership that surfaces Le Figaro journalism in ChatGPT through summaries and excerpts, with links back to the original stories. Source

6. OpenAI’s Head of Strategic Futures Publishes “The Silent Consensus” on Keeping Control of Institutions as AI Advances

OpenAI published an essay by Dean Ball, its Head of Strategic Futures, on how more capable AI could reshape society and how people can keep control of their institutions. Source

7. OpenAI Gives Four Independent Auditors On-Site Access to Unreleased Models, Source Code and Training Data

OpenAI said Apollo Research, METR, Goodfire and Transluce will work from its offices with access to unreleased models, source code, training data and monitoring systems, depending on assignment. An independent safety board receives unredacted reports. Auditors publish independently, but OpenAI can redact confidential material and delay reports so risks can be fixed first. OpenAI and Transluce plan an eight-week follow-up in early 2027. Source