Nvidia Probes Its Own CEO and Goldman Locks Up Claude Code -- AI Brief November 9

Good day %%first_name%%. Nvidia’s board has opened an independent investigation of its own CEO after reports that he was the target of a Chinese intelligence operation, and Jensen Huang has admitted to an affair in an email to staff. Goldman Sachs has taken Claude Code away from most of its traders and bankers. Also today: a Philadelphia banner with a Chicago skyscraper in it, Dr. Dre and Ice Cube disagreeing about AI, and a developer who walked away from Claude Code. One more thing before we start: the Museum of TEENY Art, a vending machine that sells collectible mini prints, opened November 3 at Here BK in Brooklyn, and its first show is on view through December 1. Go say hi.

Nvidia’s Board Opens a Probe of Its CEO NPR

* What happened: Nvidia’s board said Sunday that a committee of independent directors, assisted by outside counsel, will investigate reports that CEO Jensen Huang was the target of a foreign intelligence effort, according to Nvidia’s statement. The New York Times alleges that a Chinese businesswoman named Alice Huang, who is not related to him, built a close relationship with him. Axios reports that Huang admitted in a message to staff that he was unfaithful to his wife and said he will cooperate fully with the investigation.

* Why it matters: Nvidia dominates the market for AI chips, so who buys them is a national-security question. The Times reports that Alice Huang’s employer became one of Nvidia’s largest buyers, at about $4.8 billion in equipment, according to Al Jazeera, and NPR says the reporting suggests she may have helped China get around US limits on exports of advanced AI chips.

* What everyone’s saying: The board says it has “no reason to believe that either national security or NVIDIA secrets were compromised,” and Nvidia says Huang will keep running the company during the probe. Three House Republicans asked him to appear before Congress on November 18, Al Jazeera reports, and Senator Elizabeth Warren said, “NVIDIA’s Jensen Huang and NVIDIA’s board owe Congress answers.”

* My read between the lines: The affair is the headline. The $4.8 billion is the story. The open question is not who Huang had dinner with, it is whether anything moved through that relationship to a buyer of that size, and nobody outside the board’s inquiry has shown that it did. Everything here beyond Huang’s own email is the Times’s reporting and the board’s reply, so treat the national-security part as alleged until the investigation says otherwise.

📖 Further reading: The US Government Just Took Anthropic's Best AI Model Offline — Here's Why — my June read on the first time Washington put a frontier AI model behind a national-security gate, and the part nobody was saying out loud.

When something goes sideways at Nvidia, the board brings in outside counsel. When something goes sideways at your company, it’s usually a contract you signed at 4:55 on a Friday. Marblism sells AI employees that work inside the tools you already use, and Linda is the legal assistant who reviews contracts. No prompting skills needed: describe your business and goals, and the employees check in when they need your approval. New users get 10% off for life through this link, and there’s a 7-day money-back guarantee. Meet Linda →

Goldman Puts Claude Code Behind an Engineers-Only Rope Bloomberg, via FA Magazine

* What happened: Goldman Sachs updated its internal guidance in October to limit Anthropic’s Claude Code and Microsoft’s GitHub to engineers, Bloomberg News reports. Traders and bankers had been using the tools to monitor client trading flows, model market moves and positions, and produce written and graphic material. “Strats,” the staff who sit between trading, quant research and software, still have access.

* Why it matters: As recently as July, Goldman’s chief information officer, Marco Argenti, was praising “vibe coding” by non-technical staff, who passed their creations to the bank’s roughly 12,000 engineers to improve. State of AI points out that Claude Code reads files, writes programs and runs them with the access of whoever is using it, which is a different risk on a trading desk than inside an engineering review process.

* What everyone’s saying: A Goldman spokesperson said the bank puts these tools in the hands of “those who are best suited” to use them for the long term after a period of experimentation. State of AI’s take is that Goldman was right to restrict the tools and wrong in how it did it: a guideline update with no replacement tells the traders who found good uses that they were the risk.

* My read between the lines: Bloomberg also notes that Goldman is lead underwriter on Anthropic’s IPO, so the bankers who will spend weeks selling the company are mostly locked out of its c