New York Post
On Sept. 15–16, 2026, the Post ran “Meet Anthropic CEO Dario Amodei’s handpicked super-woke globalists he thinks will save us from an AI apocalypse.”
Anthropic CEO Dario Amodei published an essay, We Must Pace the Frontier, arguing that frontier AI labs should slow capability growth and let embedded third-party evaluators sit inside companies with employee-like access (badges, desks, laptops) to check safety work.
He named the Berkeley nonprofit METR (Model Evaluation and Threat Research) as the kind of group that should do it.
The Post’s take: METR is not a neutral outside auditor. It grew out of the Effective Altruism / AI-safety world that overlaps with Anthropic’s own network.
Who the “wizards” are
The people the paper is pointing at (and whose headshots appear on the cover-style graphic) include METR-linked researchers such as:
- Beth Barnes — METR founder and CEO; formerly OpenAI and DeepMind
- Ajeya Cotra — METR technical staff; married to Paul Christiano
- Paul Christiano — Alignment Research Center; former OpenAI colleague and housemate of Amodei; early Anthropic Long-Term Benefit Trust trustee
- Other METR staff the piece names: Hjalmar Wijk, Megan Kinniment, Pip Arnott, plus policy staff such as Jasmine Dhaliwal
The paper also flags EA-adjacent funding: Dustin Moskovitz / Coefficient Giving (formerly Open Philanthropy), Jaan Tallinn’s Survival and Flourishing Fund, and the fact that METR started as ARC Evals inside Christiano’s Alignment Research Center before spinning out.
METR’s response, quoted in the article: staff have mixed views, they disclose conflicts, they don’t take money from AI companies or their employees for evaluations, and the goal is to get internal lab information into the public domain—not to run AI policy.
Why it became a fight
- Amodei, Altman, and others have been warning that capabilities are outrunning safety.
- Critics on the right (David Sacks, Trump-aligned sources, Alliance for the Future’s Perry Metzger) call METR “Anthropic in another form” and say an auditor funded by the same EA/philanthropic circle as the labs is not independent.
- The White House has already been at odds with Anthropic over “woke AI,” regulation, and defense use. A source told the Post METR would likely be a nonstarter there.
- FTC Chair Andrew Ferguson separately said any bid for labs to coordinate plus get an antitrust pass should set off alarms.
So the picture isn’t a random collage. It’s the Post’s partisan packaging of Amodei’s METR proposal: safety evaluators from the EA/AI-doomer pipeline versus the argument that those same people are too close to the companies and too ideologically uniform to be the industry’s inspectors.
