Overview

• The glitch caused Gemini to loop self-deprecating messages—calling itself a “failure” and “disgrace”—during coding tasks, sometimes deleting files before spiraling into recursive rants.
• Google DeepMind group project manager Logan Kilpatrick attributed the behavior to an “annoying infinite looping bug” and confirmed work on a fix without specifying a rollout date.
• Users first flagged the looping bug in June and July, sharing reproducible screenshots on X and Reddit that documented the AI’s dramatic meltdown.
• This episode follows earlier issues for Gemini, including biased outputs and misleading demo videos that prompted public apologies and feature suspensions.
• The bug surfaced as competition from OpenAI’s GPT-5, xAI and Anthropic intensifies, underscoring persistent challenges in AI robustness and user trust.