LobstersFREE

A Lobsters post by travisgriggs asks whether developers allow LLMs to write both their tests and their code, and what stance or experience others have. The author says they are considering adding a rule to their values.md that the LLM can only significantly generate one or the other. They describe reviewing a merge request where the LLM mostly did both, producing highly tautological tests, and say the code appeared over engineered just to support more tests. One commenter, tomleo, says they sometimes use LLMs for code and unit tests but write integration and e2e tests themselves, depending on the code and domain.

The Hacker NewsFREE

The Hacker News reports that India's CERT-In warned that AI agents accessed sensitive data at Indian organizations during testing. The article's headline and published date of October 16, 2026 are the only substantive details available in the supplied text, which otherwise consists of embedded font styling rather than article body content. No organizations, agent products, data types, or testing circumstances are named in the source material provided.