Goldman Sachs 将 Claude Code 收紧至约 12000 名工程师使用

这一决定来自对生产力和风险的内部评估：全员开放后使用量猛增，但实际效率提升难以衡量，交易部门一条错误脚本可能造成数百万美元损失。无独有偶，Apple 上周也限制了工程师的 AI 支出额度。值得注意的是，Goldman 同时是 Anthropic IPO 的主承销行之一，两个角色并存颇为微妙。

HUGE news. One side Goldman is flogging Anthropic’s IPO while the other is taking away their tools for many of their own employees. Hedgie @HedgieMarkets 🦔Goldman Sachs took AI coding tools away from most of its traders and bankers. New internal guidance limits Anthropic’s Claude Code and Microsoft’s GitHub tools to Goldman’s roughly 12,000 engineers, after a review of productivity and risk. Traders had used them to track client flows and model market moves. Back in July, Goldman’s tech chief publicly praised what traders built with AI. My Take I’d have made the same call. Handing AI to everyone sent usage through the roof. Whether anyone got faster is a lot murkier, and on a trading desk one bad script can cost millions. A lot of Anthropic’s growth came from companies that put Claude on every desk because the boss said so. Apple capped its engineers’ AI spending last week. Goldman is also one of the lead banks taking Anthropic public, which should make for some awkward pitch meetings. I believe we'll see more companies cut back as the bills come in, and Anthropic’s growth to slow right along with them. Hedgie🤗 🔗 View Quoted Tweet📊 3 ⚡

This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Anthropic, Claude Code, Goldman Sachs, GitHub.

Source attribution: curated from Gary Marcus (@GaryMarcus).