标签： Claude

Q&A with Mustafa Suleyman on recent AI safety incidents, risks of removing guardrails while testing 10x-larger future models, a cross-industry safety body, more (Shirin Ghaffary/Bloomberg)

Q&A with Mustafa Suleyman on recent AI safety incidents, risks of removing guardrails while testing 10x-larger future models, a cross-industry safety body, more (Shirin Ghaffary/Bloomberg)

微软 AI 负责人 Mustafa Suleyman 就近期多起 AI 安全事件接受采访，谈到在测试比现有模型大 10 倍以上的未来模型时移除安全护栏的风险，并再次呼吁建立跨行业的安全机构。核心矛盾是：模型能力还在加速，但安全治理的组织形式仍未定型。

(4) Katie Notopoulos on X: “This Makes Me Feel Insane! This Is Faux-Hallmark Pap and if You Believe for 1 Sec That He Cares About Helping People “spending More Time With Family” or “opening a Bakery” You Don’t Need to Wait for AI to Kill Us; You’re Stupid Enough to Drown From Looking Up at the Rain” / X

(4) Katie Notopoulos on X: “This Makes Me Feel Insane! This Is Faux-Hallmark Pap and if You Believe for 1 Sec That He Cares About Helping People “spending More Time With Family” or “opening a Bakery” You Don’t Need to Wait for AI to Kill Us; You’re Stupid Enough to Drown From Looking Up at the Rain” / X

Scale AI 创始人 Alexandr Wang 发表《Why We're Building Muse》一文，宣称 AI 会替人解决所有问题、只留下"渴望与梦想"，招致科技作者 Katie Notopoulos 的激烈批评。这场争论的焦点不是技术路线，而是 AI 叙事中"解放人类"的说法是否可信。

How AI’s acceleration created a global policy vacuum, as EU AI Act enforcement lags and regulators remain torn between harnessing AI and fearing its risks (New York Times)

How AI's acceleration created a global policy vacuum, as EU AI Act enforcement lags and regulators remain torn between harnessing AI and fearing its risks (New York Times)

据《纽约时报》报道，AI 能力迭代速度已明显快于监管落地节奏，欧盟《人工智能法案》的执行进度落后，全球正在形成一段政策真空期。监管者一方面想借 AI 提振产业竞争力，另一方面又担心模型失控与系统性风险，这种摇摆正在直接影响 AI 产品能否上线、以什么条件上线。

Show HN：TinyAIArena 看 AI Agent 一决高下

Show HN：TinyAIArena 看 AI Agent 一决高下

开发者 hp6 在 Hacker News 上发布了 TinyAIArena，让多个 AI Agent 在一个回合制对战游戏里互相对打，任何观众都能点开比赛实时围观。它值得关注的地方不在于游戏本身，而在于这是一次用博弈环境测大模型决策能力的公开实验。