AI Reading List

Daily curated list of news, articles, and other interesting posts related to AI.

2025-11-08

• Oddest ChatGPT leaks yet: Cringey chat logs found in Google Analytics tool (arstechnica.com)

  Strange ChatGPT conversation logs containing sensitive and personal information have appeared in Google Search Console, raising concerns about privacy and data handling. Experts suggest this may be evidence that OpenAI scrapes Google searches, uncovering unexpected data exposure for ChatGPT users.

• AI benchmarks are a bad joke – and LLM makers are the ones laughing (www.theregister.com)

  Interesting take on AI benchmarks, highlighting that many current tests do not clearly define what they measure and often lack rigorous scientific methods. The article points out that claims about model superiority, such as those made for GPT-5, may not be as meaningful as they appear due to these flaws.

• Firefox Forcing LLM Features (equk.co.uk)

  Overview of concerns about Firefox introducing AI and LLM features by default, with limited options for users to disable them and reports of increased resource usage. The post includes a guide to turning off these features and mentions alternatives for privacy-focused users.

• How to declutter, quiet down, and take the AI out of Windows 11 25H2 (arstechnica.com)

  Practical guide to removing unnecessary features and AI components from a fresh Windows 11 25H2 installation, aimed at users upgrading from Windows 10 or performing clean installs. Highlights steps for streamlining the operating system and minimizing built-in AI integration.