Gemini 3 Shows Autonomous Programming Testing Abilities, Sparking Discussion on New AI Capabilities

On the Linux.do tech forum, a user first discovered that Gemini 3 claimed in its responses to have written and verified code through testing, a phenomenon that has sparked heated discussion. Users speculate this might be a newly introduced feature, suggesting potential breakthroughs in AI model autonomy and code generation capabilities. The discussion spans multiple posts, with participants analyzing this observation and exploring its impact on AI development workflows. This discovery not only reflects the rapid evolution of AI technology but may also offer new approaches for automated testing and code optimization. Although details are limited, as a cutting-edge technology development, it deserves attention from tech enthusiasts, especially users interested in the evolution of AI autonomy.