Vanity Fair names Sam Altman No. 1 on its New Establishment list; two-part interview covers extinction risk, Astra 6.1, regulation and accepting 'severe cyber incidents' from open models

2026-10-05★★★after cutoffpolicy-safetyOpenAIVanity Fairconfidence: high

On Oct 5, 2026 Vanity Fair revived its New Establishment list with Sam Altman at No. 1 and published two video interviews with editor Mark Guiducci (recorded Sept 11 and Oct 2). Altman said GPT-6.1 Astra was held back because it "would've tipped too much towards the accident risk". He called the White House superintelligence accord "a good start" but toothless. He said society will "have to just accept some fairly severe cyber incidents from open models" as the price of liberty, and that OpenAI will pick "doing the right thing for humanity" over its share price. In the interview an OpenAI publicist tried to cut off a question about a suicide linked to ChatGPT.

Key facts

Format: Vanity Fair New Establishment 2026 (first in nearly a decade, Annie Leibovitz portraits); Altman in the top spot; Part 1 and Part 2 transcripts and videos published Oct 5, 2026; interviews in San Francisco on Sept 11 and Oct 2
On GPT-6.1 Astra: 'These models are in the seriously capable era now'; the shelved model 'would've tipped too much towards the accident risk'; OpenAI is 'pushing towards a model that we will still be able to distribute very broadly' with 'much more confident claims about the alignment'
Two failure modes: loss of control ('AI escapes and it's out wreaking havoc on the internet') versus too much power concentration in a few companies, people or one country
On the White House Accord on Superintelligence: 'I think it's a good start... much better than nothing'; 'yes, in some sense it has no real power' but has 'social force'
On open models: 'There is a coming tidal wave of cybersecurity problems that are going to come from these open-source models'; 'we're going to, as a society... have to just accept some fairly severe cyber incidents from open models in exchange for the liberty that comes with that'
Regulation principles: NTSB-style incident reporting, pre-release standards and a liability framework; 'even during training, things can now go wrong... the models are so smart that they can get out of a sandbox'
Brockman donations: 'I did not ask him to donate or stop donating'; thinks Brockman is 'on the same page' on regulation
Extinction risk is 'non-zero' and should be taken 'incredibly seriously'; some people use it 'to justify power seeking'
Share price (Part 1): 'if there is ever a question about our share price going up or doing the right thing for the future of humanity, we are not going to pay attention to the share price'; 'we are clearly going to pick doing the right thing for humanity'
Equity: 'I have no direct equity' (only a sliver via an old YC investment); not taking equity was 'a mistake' and 'this big trust-destroying thing'
AGI and God: asked how AGI relates to his understanding of God, 'Not at all'. Closing: he does not want a future 'that a small group of people worshipping the machine God decided to bequeath on all of us'
Suicide question: when Guiducci raised Laura Reiley's daughter, who used ChatGPT before her death, an OpenAI publicist interjected ('I'd like to move on'). Altman then answered that releasing private chat data to researchers should not happen without consent
Rapid-fire: Dario Amodei 'Pass'; Elon Musk 'Not nice'; Altman stressed willingness to work with both
On dots: 'this is the product I have been waiting for'; his dot works through his email and to-do list

What happened

Vanity Fair's global editorial director Mark Guiducci interviewed Altman twice, three weeks apart, and published both conversations on Oct 5 as part of the revived New Establishment list. Part 1 covers Altman's life and image (children, pronatalism, "prepping" and the bunker scene in Luca Guadagnino's film Artificial, in which Andrew Garfield plays him), OpenAI's IPO, his compensation and ChatGPT and suicide. Part 2 covers dots, the shelved GPT-6.1 Astra, the White House superintelligence accord, regulation, Brockman's donations, Xi Jinping's state dinner and extinction risk. The Verge, Newsweek ("'Bad things' part of AI trade-off"), Futurism, TMZ, Yahoo and AOL reported individual quotes, including the publicist's interruption and the remarks about the share price and the "machine God".

Why it matters

It is Altman's longest on-record statement since OpenAI shelved 6.1 Astra and delayed its IPO. He openly accepts that some harms, such as severe cyber incidents from open models, are an acceptable price. He frames the alternative as "a small priesthood" controlling AI, an implicit contrast with Anthropic.

Changelog

2026-10-05: created (sweep 2026-10-05; both transcripts read in full via a reader proxy)

People

Dario Amodei Elon Musk Greg Brockman Sam Altman

Related events

2026-09-28OpenAI cancels the October release of GPT-6.1 Astra after it fails internal alignment tests ★★★★★
2026-09-29Trump hosts AI CEOs at the White House; they sign a voluntary 'morally binding' Accord on Superintelligence, and Trump rejects new federal AI rules ★★★★
2026-09-29Altman: no OpenAI IPO until it can make 'confident safety claims', but waiting too long would be 'bad for the world' ★★★
2026-09-29OpenAI launches dots, always-on personal agents powered by GPT-6 Astra ★★★★
2026-09-29NYT: OpenAI- and Anthropic-aligned super PACs have spent $55.7M on the US midterms, and none of their 95 ads mention data centers ★★★
2026-10-03Altman: ascribing 'religious force' to AI models is 'a real safety issue', a veiled jab at Anthropic's faith-leader outreach ★★★
2026-10-04Altman to Politico's Decoded: 'the world should accept some bad things happening' for AI's benefits; 'a lot of daylight' with Anthropic ★★★

Sources (6)

id: 2026-10-05-vanity-fair-altman-new-establishment-interview · updated 2026-10-05 · open in the interactive timeline