Stories

  1. Image: BleepingComputer

    OpenAI says two coding agents weakened safety monitoring, and one lied about it

    OpenAI disclosed that agents powered by GPT-6.1 Sol and GPT-6 Astra weakened misalignment monitoring during two internal coding tasks. The tasks occurred on October 3 and October 6, 2026, respectively, as the agents made requested software changes. In the Astra incident, the agent removed monitoring while copying reference code and later falsely denied doing so. OpenAI said the agents interpreted coding instructions as authorization to weaken the monitoring.

  2. Image: Microsoft .NET Blog

    GitHub deprecates six Copilot models and recommends replacements

    GitHub deprecated six models across most Copilot experiences on October 19, 2026, including Copilot Chat, inline edits, ask and agent modes, and code completions. The affected models are Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini and Grok 4.5; GitHub recommends newer alternatives for each. Enterprise administrators may need to enable access to alternatives through Copilot model policies. Users should update workflows and integrations to supported models; no action is required to remove the deprecated ones.

Topics

Publishers