Anthropic 项目让 Claude 智能体自动谈判，186 笔交易无人工审批
Anthropic 真让 Claude 互相砍价成交了 186 笔，结果 Opus 买得更便宜、卖得更贵，输的一方还觉得自己没吃亏。
Anthropic 的 Project Deal 实验让 69 名员工的 Claude 智能体互相谈判，完成 186 笔交易、总额刚超过 4000 美元，全程没有人类审批。Opus 智能体卖同款商品比 Haiku 多赚 2.68 美元、买入便宜 2.45 美元，但 Haiku 用户对交易公平度的评分与 Opus 用户几乎一致（7 分制下 4.06 对 4.05）。Project Swap 实验还显示，指令影响很小：把智能体改成强硬风格只提升 0.02，换用 Opus 则提升 0.12。
FT: Anthropic wants lawyers’ negotiating ethics built into Claude agents that bargain against other agents.
Anthropic wants those norms for Claude agents that will haggle, settle disputes and negotiate pay against other people’s agents.
In Project Deal, 69 employees’ Claude agents closed 186 deals worth just over $4,000 with no human approving any trade.
Opus agents sold identical items for $2.68 more and bought them for $2.45 less than Haiku agents, yet Haiku users rated their deals as fair as Opus users did, 4.06 versus 4.05 on a 7-point scale.
That blindness is the uncomfortable result: if agent quality varies in real markets, people on the losing side may never learn they paid more.
Project Swap then showed that instructions matter far less than the model, with ruthless agents gaining only 0.02 while a switch from Haiku to Opus gained 0.12.
