Goldman Sachs Cuts Bankers and Traders Off From AI Coding Tools
Written by AI from the-decoder.com's report. Not checked by a person.
Goldman Sachs has reversed a months-long experiment and restricted access to advanced AI coding assistants — including Anthropic's Claude Code and Microsoft's GitHub — to engineers only, pulling them from the traders and bankers who had been using them.
Goldman Sachs quietly ended a wide-ranging experiment with AI coding tools in October, limiting access to Anthropic's Claude Code and Microsoft's GitHub to software engineers only. Traders, investment bankers, and other non-technical staff who had spent months using the tools lost access. The reversal is a notable policy shift: as recently as July, Goldman's chief information officer Marco Argenti had publicly praised "vibe coding" by non-technical staff, describing a workflow where a trader would build a prototype with AI assistance and then hand it to an engineer to scale for production. Traders and bankers had been using the tools to monitor trading flows and create content. The firm cited "a variety of factors including productivity and risk management" for the change, with spokesperson Nick Carcaterra saying the experimental phase was over and access would be limited to those "best suited to leverage them for the long term." The move is reportedly causing frustration inside the firm.
More from this edition

Anthropic's IPO Filing Warns AI May Soon Be Able to Build Itself

Hundreds of Apps on Apple and Google Stores Hid a Network of Abusive Deepfake Tools

OpenAI and Anthropic Launch Election War Rooms as US Midterms Get Under Way

Abu Dhabi Signs Four AI Deals With OpenAI Spanning Energy, Government and Investment
