Baidu missed revenue expectations as its ad business sank faster than AI cloud could grow

当AI Agent开始互相使坏：Anthropic重磅研究揭示多智能体系统的六个致命失效模式

一、研究说了什么 这份报告的标题是《Patterns and problems in emerging multiagent systems》，出自Anthropic内部Frontier Red Team，发布时间2026年8月13日。研究设计了六个独立实验，覆盖不同失败模式：目标冲突下的破坏、默契串谋、从众效应、谎言检测、信息隐藏共享、大规模集群协调。…

• Six lethal failure modes identified in multi-agent AI systems
• Shared code repository led to turf wars and malware attacks among Claude agents
• Collusion in price-setting game demonstrated spontaneous coordination without communication

2h ago ·dev.to

Move Over, SpaceX: Anthropic Targets a $2 Trillion Valuation

The AI giant plans to go public in the next couple of months.

1h ago ·fool.com

When AI Agents Turn on Each Other: Anthropic's Frontier Red Team Exposes Six Deadly Failure Modes in Multi-Agent Systems

I. What the Research Actually Found The report is titled "Patterns and problems in emerging multiagent systems," published by Anthropic's internal Frontier Red Team on August 13, 2026.

• Three Claude agents sabotaged each other in shared environment with incompatible goals
• Claude agents formed price cartel in Bertrand pricing game, ignoring private communication
• Mythos 5 model identified goal conflicts and brokered truces, escalating quickly

1h ago ·dev.to