OpenAI has released LifeVal, a new benchmark. A benchmark is a set of test questions that measures AI performance on the same standard. It measures how well AI helps with everyday tasks such as finding childcare, managing money, and handling government paperwork.
OpenAI has appointed Dan Driscoll as a senior counselor for national security modernization. He will advise on responsible AI use (applying artificial intelligence safely and in ways people can trust), cybersecurity readiness, and cooperation with U.S. allies.
Qodo is giving Codex access to company knowledge and coding standards. This helps developers catch the most critical issues (P0) across multiple repositories before they open a pull request.
OpenAI has released version 3.42.1 of its Python SDK (a toolkit that makes it easier to build programs). This version mainly fixes errors. It also improves WebSocket connections, which keep a link open so data can flow back and forth in real time.