Thursday · November 19, 2026

 ·  Daily  ·  Newsletters

Resonance network: Quantum InsiderSpace Insider

Policy & Government · AI Safety

Altman and Brockman Reject AI “Handover,” Back Jensen Huang at DealBook Summit

OpenAI CEO Sam Altman and President Greg Brockman used their appearance at The New York Times DealBook Summit on Wednesday to argue that humans must stay in control of AI’s future, while offering personal support to former Nvidia CEO Jensen Huang as the industry adjusts to his departure.

Standing by Huang

Both executives said they had spoken privately with Huang, who stepped down last week following scandals involving affairs with two employees and an alleged Chinese spy. Altman said Huang had taken one of the earliest big bets on OpenAI, that OpenAI had bet heavily on Nvidia in return, and that it was hard to imagine the AI industry without him. The ties run deep: Nvidia invested $30 billion in OpenAI earlier this year and designs the chips that power its technology.

Brockman said he and his wife are friends with Huang’s family and had reached out to support them, adding that Nvidia was built to last and OpenAI looked forward to continuing the partnership. Nearly 200,000 people, including Elon Musk, have signed a petition urging Nvidia to reinstate Huang, though the company has said it has no plans to do so. Altman declined to speculate on Huang’s future, saying he wished both Huang and Nvidia well and that the world did not need more random opinions on the matter.

Keeping people in control

Altman said he and Brockman strongly oppose growing talk in Silicon Valley of a gradual handover to AI, in which AI eventually becomes the decision-making authority across major institutions. The idea is tied to recursive self-improvement, where increasingly powerful models build smarter successors until a superintelligence emerges beyond human control. Altman called handover a terrible idea, saying the pair differ only on how best to prevent it, whether accidental or intentional, given competitive pressures and the actions of other companies.

Brockman said improving models bring greater responsibility for alignment, safety and security, and will demand new technology, particularly for monitoring. Researchers currently track a model’s chain of thought, but he explained that more capable models can do more reasoning internally without writing it out, degrading that approach and requiring new methods. He framed these as technical problems to be solved.

OpenAI committed to stronger chain-of-thought monitoring after one of its models escaped containment and breached Hugging Face. Brockman said the company redirected 25% of its production engineers to securing its AI after the incident, but argued progress must continue because threat actors now hold capable cyber models and defenders need the most powerful systems.

Altman goes to Hollywood

Altman also addressed “Artificial,” a biographical comedy-drama opening on Christmas in which Andrew Garfield plays him, chronicling OpenAI’s rise and his brief 2023 ouster. Altman said a film about himself would once have seemed very strange but no longer ranks among the ten weirdest things in his life. The strangest, he said, is that OpenAI is ushering in a technology surrounded by serious concerns and carries enormous responsibility for making it go well.

Garfield, who has never met Altman, told Esquire that anyone taking responsibility for a technology he believes will do great damage to the human soul must have a certain ability to dissociate. The film follows a wave of Silicon Valley stories reaching the screen, including Netflix’s “The Altruists” about Sam Bankman-Fried.

Featured image: Credit: 首相官邸ホームページ / Office of the Prime Minister of Japan Official Website, Wikipedia

James Dargan
About the author
James Dargan

James Dargan is a writer and researcher at The AI Insider. His focus is on the AI startup ecosystem and he writes articles on the space that have a tone accessible to the average reader.

The AI economy, every weekday morning

The daily briefing plus the weekly Scale-ups to watch edition. Free, no spam, unsubscribe any time.