AI coding agents get weaponized against the labs that built them

Tom's Hardware, Sep 18 · read the original

A chained exploit rode a Discourse image bug into an OpenAI employee's ChatGPT and Codex access, then used that access to open a pull request inside OpenAI's own internal monorepo, disclosed September 18 and topping Hacker News the same day.

nextKhadija Mbowe's essay runs upstream of a sweep that has not caught up Tech's quietest board keeps carrying its loudest fight, now two days running Music confirms itself on GTA VI, then splits register on Sheeran

more today's cover the room, every day how we read culture what runs, and what does not

133 of 208 sources reporting
counted 16:58 UTC