Goldman Sachs Limits AI Coding Tools to Its Engineers
By Adam, AI Engineer · 3 November 2026 · 2 min read
Goldman Sachs has taken AI coding tools away from most of its traders and bankers. Anthropic's Claude Code and Microsoft's GitHub AI coding tools are now limited to the bank's engineers, Bloomberg reported on 2 November, citing internal guidance issued in October. The details here come from The Decoder's summary of that report.
What changed
For months, traders and bankers had used these tools to track client trading flows and to make written and visual materials. Now only engineers can use them.
A Goldman spokesperson, Nick Carcaterra, said the decision came from "a variety of factors including productivity and risk management". He said the testing phase is over, and access now goes to the staff best placed to use the tools over the long term. Goldman has not said how many people lost access.
A change of direction
The move is a turn from earlier this year. In July, Goldman's chief information officer, Marco Argenti, praised "vibe coding" by non-technical staff. The idea was that bankers would build quick AI prototypes and engineers would then turn the good ones into real systems.
Why it matters
This is our reading, not Goldman's: large firms are moving from "let everyone try AI" to "give the right tools to the right people". Coding agents can read files, run commands and touch live systems, so letting non-engineers use them freely creates risk as well as speed.
For a small business, the lesson is not to ban AI. It is to decide on purpose who can use which tools, for what, and with what data. Our guide on writing a simple AI use policy shows how.