OpenAI Staff Defend Sam Altman’s Openness to Criticism

A wave of OpenAI employees and former staffers took to X to publicly defend CEO Sam Altman's leadership style, specifically his willingness to hear criticism from inside the company. The posts came in response to a since-deleted message from Nick Huber, described as an AI product leader, who wrote about a negative interview experience at OpenAI. Current researchers and two ex-members of the Sora team pushed back on any suggestion that the company punishes honest internal dissent.

What happened

Nick Huber, described as an AI product leader, posted on X about what he characterised as a poor interview experience at OpenAI. Screenshots of the post indicate that an interviewer asked him to use data to challenge a belief attributed to Altman. The post has since been deleted. OpenAI did not confirm to Business Insider whether Huber had been interviewed, and Huber did not respond to a request for comment.

That post prompted a string of public responses from OpenAI insiders defending the company’s internal culture.

Who said what

Eric Mitchell, who co-leads OpenAI’s Post-training Frontiers team, wrote that he had “directly disagreed with, corrected, or expressed frustration with leadership to Sam” on multiple occasions, including during his first six months at the company. He described Altman’s response as consistently curious and open-minded, and said claims of a retaliation culture were inaccurate.

Researcher Brandon McKinzie added that he has delivered critical feedback directly to Altman and other leaders, and that “not only do they listen to critical feedback, they take it very seriously and take action on it.” He called it “one of OpenAI’s greatest strengths.”

Two former members of the Sora team also weighed in:

• Gabriel Petersson, previously a researcher at OpenAI and Midjourney, now running his own company: “Can confirm, Sam is super receptive.”
• Will DePue, who left OpenAI in April: described the company as “exceptionally receptive to internal disagreement and critique, for better or worse.”

Victor Nunez, who works on OpenAI’s Codex coding tool, described “memes and hard conversations on a daily basis” and said the culture was “as human as it gets.”

Why it matters

Altman’s leadership has faced public scrutiny since November 2023, when OpenAI’s board briefly removed him as CEO. At the time, the board stated he was “not consistently candid in his communications.” Altman, during a recent court dispute with Elon Musk, said: “I believe I am an honest and trustworthy business person.”

Employees have referred to the 2023 episode as “the Blip.” It is now the subject of an upcoming film directed by Luca Guadagnino, which was picked up by Neon after Amazon dropped the project following its deepening commercial relationship with OpenAI. The story of what really happens inside frontier AI labs, including how dissent is handled, matters for anyone evaluating whether to build on or integrate these platforms. A lab with a fragile feedback culture is more likely to ship models with unchecked blind spots.

A 2021 study cited in the Business Insider report found that open criticism sharing in workplaces increases employees’ sense of psychological safety. Google separately listed psychological safety as the top trait of effective teams in 2015.

For businesses that are integrating AI tools from labs like OpenAI, the internal culture of those labs is not just HR trivia. It shapes what gets flagged before a model ships and what gets quietly ignored.

Our take

A coordinated wave of positive posts from current employees defending a CEO, triggered by a deleted complaint, is a PR pattern worth noting. That does not mean the staffers are wrong. Mitchell’s account is specific and credible: he names a timeframe, describes repeated interactions, and gives detail that goes beyond “my boss is great.” DePue’s qualifier “for better or worse” also adds a note of honesty that generic praise usually lacks.

That said, the people most likely to speak up are the ones who had good experiences. The deleted original post and OpenAI’s non-response leave an obvious gap. If you are a developer or business owner deciding how much to depend on OpenAI’s products, this episode tells you the company has staff willing to publicly back its culture. It tells you less about whether that culture is uniform.

For context on how other AI labs handle talent and internal dynamics, see our coverage of why experienced tech professionals are joining AI labs despite already being wealthy.

What to do about it

1. If you are evaluating which AI platform to build on, look past the PR posts and check the lab’s model cards, safety reports, and public incident logs for concrete evidence of internal quality control.
2. Watch how OpenAI handles the next major model controversy, not this one, since that will be a cleaner signal of whether the feedback culture holds under commercial pressure.
3. If you are deploying OpenAI tools at scale, diversify your integrations enough that a single lab’s cultural or product shift does not strand your workflow.

The most useful thing a company can do to prove an open feedback culture is to act on feedback publicly, not just defend it privately.

Source: Bing News · Sora (AI video)