AIFortuneabout 22 hours ago
Alibaba.com President: AI agents need outcome tests
6 min read
Alibaba.com president Kuo Zhang says AI agents must be graded on completed work, not conversation. The Accio team's open-source CommerceAgentBench tested 107 real e-commerce tasks; the strongest frontier model completed only 61.7%. No single model won across all categories.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
