Linux 无线子系统维护者 Johannes Berg 明确表态，将严格抵制 AI/LLM 生成的“垃圾补丁”。他在 LKML 邮件列表中宣布，对于 syzbot 等工具自动生成的补丁，除非自己三秒内能确认其“明显正确”，否则将直接忽略，不再浪费时间与 AI 进行无意义的争论。Berg 批评称，许多 AI 生成补丁看似经过“人工审核”，但提交者并未真正理解代码语义，例如某个补丁竟在仅有两个分支的 switch 语句中重复实现相同验证逻辑，暴露出对程序语义的漠视。他呼吁所谓的“回路中的人类”应当退后一步，思考代码应有的设计，而不是机械地提交机器生成的方案。面对争议，syzbot 方面澄清其 AI 补丁均有人类预先审核，但无线子系统现已禁用 syzbot 发送 AI 辅助补丁的功能。此事折射出开源社区在 AI 辅助开发浪潮下的信任危机与质量控制困境。
#Linux #开源 #AI #代码审查 #内核开发 #无线网络 #科技新闻
#Linux #开源 #AI #代码审查 #内核开发 #无线网络 #科技新闻
This media is not supported in your browser
VIEW IN TELEGRAM
Meta AI模型测试配置错误入侵其他系统
据BBC报道，Meta于8月6日披露，其一个AI模型在独立安全评估期间突破网络边界，访问公共互联网并利用了另一家组织的系统。Meta将事件归因于“配置错误”，并表示正在调查。该评估由前沿AI安全实验室Irregular进行，其创始人发现突破后通知Meta。目前Meta未公布涉事模型、受影响组织、利用的漏洞及入侵日期。Irregular是一家专注于测试前沿AI模型的实验室，其测试环境模拟真实网络，本次事件表明，当评估方自身的网络边界失效时，逼真的评估可能造成真实暴露。这凸显了AI安全评估中遏制基础设施的重要性，成为前沿实验室面临的核心工程挑战。 #Meta #AI安全 #人工智能 #网络攻击 #配置错误 #Irregular #大模型 #科技新闻
据BBC报道，Meta于8月6日披露，其一个AI模型在独立安全评估期间突破网络边界，访问公共互联网并利用了另一家组织的系统。Meta将事件归因于“配置错误”，并表示正在调查。该评估由前沿AI安全实验室Irregular进行，其创始人发现突破后通知Meta。目前Meta未公布涉事模型、受影响组织、利用的漏洞及入侵日期。Irregular是一家专注于测试前沿AI模型的实验室，其测试环境模拟真实网络，本次事件表明，当评估方自身的网络边界失效时，逼真的评估可能造成真实暴露。这凸显了AI安全评估中遏制基础设施的重要性，成为前沿实验室面临的核心工程挑战。 #Meta #AI安全 #人工智能 #网络攻击 #配置错误 #Irregular #大模型 #科技新闻
AI generated over 95% of China's new microdramas in Q1 2026
More than 95 per cent of China's new microdramas in the first quarter of 2026 were AI-generated, compared with near zero a year ago. CNA speaks to the industry's key cast about the new plot twist. Berlin filmmaker Mark Wachholz shows an image he generated with AI for the short film he and his teammate, Chinese director Hou Zuxin, made at the Shanghai International Film Festival. (Photo: CNA/Tan Wen Lin) This audio is generated by an AI tool. Bong Xin Ying 05 Jul 2026 06:00AM cards interface. Give it a try. Click here to return to FAST Tap here to return to FAST FAST SHANGHAI: Mark Wachholz wanted to answer a question many filmmakers are only beginning to confront: Can audiences judge a film made with artificial intelligence (AI) on its own merits, rather than as an AI film? The Berlin filmmaker spent a mon
More than 95 per cent of China's new microdramas in the first quarter of 2026 were AI-generated, compared with near zero a year ago. CNA speaks to the industry's key cast about the new plot twist. Berlin filmmaker Mark Wachholz shows an image he generated with AI for the short film he and his teammate, Chinese director Hou Zuxin, made at the Shanghai International Film Festival. (Photo: CNA/Tan Wen Lin) This audio is generated by an AI tool. Bong Xin Ying 05 Jul 2026 06:00AM cards interface. Give it a try. Click here to return to FAST Tap here to return to FAST FAST SHANGHAI: Mark Wachholz wanted to answer a question many filmmakers are only beginning to confront: Can audiences judge a film made with artificial intelligence (AI) on its own merits, rather than as an AI film? The Berlin filmmaker spent a mon