The attack on Sam Altman's home was an act of violence, full stop. No journalist or critic wants their work to end with a Molotov cocktail.
But condemning that attack doesn't require softening the press. Incendiary rhetoric and adversarial journalism aren't the same thing. One inflames. The other interrogates people who hold power that most of us can't meaningfully contest.
Altman himself has called AI "the most powerful tool" in human history and warned that its power "cannot be too concentrated." He's described the internal competition for AGI control as "Shakespearean," driven by a "totalizing philosophy." If the stakes are civilizational, as he says, then the culture, stability, and decision-making at OpenAI are matters of public consequence. The New Yorker's recent investigation by Ronan Farrow makes clear why.
The piece details how Ilya Sutskever, OpenAI's chief scientist, compiled seventy pages of Slack messages and HR documents alleging that Altman had a "consistent pattern of lying" to executives and board members. It describes how Dario Amodei spent years taking detailed notes under the heading "My Experience with OpenAI," eventually concluding that "the problem with OpenAI is Sam himself." It documents that the board's independent investigation produced no written report, that board members received only oral briefings, and that at least some board members were apparently never briefed at all. One person close to the inquiry described the process as designed to "acquit him." Those aren't the ramblings of disgruntled former employees. They're sourced, documented, and on the record.
Altman admits he was conflict-averse during the board crisis and made mistakes in his dealings with Elon Musk. In most contexts, that's self-awareness worth crediting. In the context of AGI development, those mistakes carry weight beyond any one company's org chart. The New Yorker reports that when the board pressed him to acknowledge a pattern of deception, he reportedly said, "I can't change my personality." A board member's interpretation: "What it meant was 'I have this trait where I lie to people, and I'm not going to stop.'" Celebrating an "insane trajectory" of success while glossing over the human fallibility steering it isn't journalism. It's a press release.
This matters because the decisions being made aren't abstract. The reporting describes Altman quietly pursuing billions in investment from Gulf autocracies while publicly welcoming AI regulation, then lobbying against it. It describes OpenAI subpoenaing a twenty-nine-year-old lawyer who helped draft a California AI safety bill, demanding all his private communications. It describes the superalignment team, announced with a commitment of 20% of the company's compute, receiving somewhere between 1% and 2% in practice on the oldest available hardware. It describes safety culture being dissolved team by team, until an OpenAI representative, asked to arrange interviews with researchers working on existential safety, responded: "What do you mean by 'existential safety'? That's not, like, a thing."
There's also a practical case for hard reporting. When critical coverage disappears, conspiracy theories fill the gap. The New Yorker piece itself is evidence of this: it documents how intermediaries connected to Musk circulated dozens of pages of opposition research containing genuinely false and homophobic smears. The reporters spent months investigating those claims and found no evidence for them. That's what journalism does. It distinguishes documented concern from bad-faith attack. A press that can't make that distinction, or that abandons the field entirely to avoid being called incendiary, leaves the public with only the smears.
We can hold both positions: unequivocal condemnation of the violence directed at Altman and his family, and an equally unequivocal defense of journalists who ask questions powerful people prefer not to answer. Altman once said that anyone with sole control over AGI represented an "AGI dictatorship," a scenario OpenAI's founders considered among the worst possible outcomes. The press exists, in part, to ensure that judgment doesn't quietly get revised away. Given what's been reported, that work is more necessary than ever.