Peter Thiel's startup Objection, which launched this spring with the goal of dragging reporters into AI-run arbitrations funded by aggrieved article subjects, has renamed itself The Primary and abandoned its tribunal model. The company now ranks journalists and outlets on a public scoreboard, with scores determined by a large language model's assessment of reporting rigor.
The pivot was detailed in a white paper and in an email to The Hollywood Reporter from CEO Aron D'Souza, the provocateur behind the Enhanced Games and the legal strategy that secretly bankrolled Hulk Hogan's privacy lawsuit against Gawker. D'Souza said the shift grew out of the recognition that verdicts punish failure but don't fix the underlying incentives. "What we learned is that a verdict is an enforcement instrument, and enforcement only changes behaviour at the margin of getting caught," he wrote.
The mechanics of the scoreboard
The Primary's system evaluates journalists and outlets on metrics including source attribution, coverage tone, and right of reply. The rankings are built from a small initial pool, with Jeff Bezos' The Washington Post and David Ellison's CBS News currently leading outlets. The Daily Mail tabloid received the lowest aggregate score. Among individuals, Reuters aviation correspondent Rajesh Kumar Singh holds the top spot, while New York Post columnist Cindy Adams and Daily Mail's Mark Halperin are among the lowest-ranked. The New York Times' AI beat reporters Mike Isaac and Cade Metz also appear near the bottom. The Hollywood Reporter and its staff have not yet been scored.
D'Souza insisted the scorecard will have a positive effect on the profession. "A reporter who's spent fifteen years doing the slower, harder, better-sourced work has never had a portable, public record proving it - something they can carry to an editor, a proprietor, or a competitor," he said. "Now they do. That's how incentives move: not by shaming anyone, but by making the alternative legible enough that someone can finally pay for it."
From enforcement to measurement
Objection's first and only tribunal, in April, targeted The Hollywood Reporter over its 2021 coverage of a Sackler family heir. Before a verdict was issued, the company concluded that adjudication alone couldn't address what it saw as journalism's root problem. D'Souza and CTO Kyle Grant-Talbot, who met at Oxford and previously founded a laundry service, acknowledged the difficulty of building proprietary software that wouldn't be dismissed as "slop justice." The Primary's home page now states that "verdicts punish failure. They don't fix the incentive."
D'Souza framed the change as a continuation of a question he's pursued since the Gawker litigation: how journalism gets evaluated. "Objection was one attempt at that question - the enforcement attempt. Primary is the second attempt at the same question, and I think it's the right one," he wrote. He described watching an AI system compress and serve a TechCrunch interview he'd given within hours, noting that "the reporter did the work. The machine took the value." No arbitration process, he said, can repair that dynamic.
A crowded field of media watchdogs
The Primary enters a space where several efforts to rate journalistic quality have already failed. Non-profit NewsTrust, which used staffers to rate stories, launched two decades ago. The automated browser extension The Factual arrived in 2016, followed by Credder in 2019, which styled itself as a Rotten Tomatoes for journalism. None remain in operation. The Primary's business is "pre-revenue," with D'Souza saying its revenue model is "a decision in front of us."
As AI systems increasingly parse and rank written work, some professionals are exploring AI for Writers Courses to understand how these models evaluate content. The Primary's methodology, while still unproven, reflects a broader push to quantify reporting quality in ways that could influence hiring and pay.
Why this matters for writers
The Primary's ranking system, if it gains traction, could create a new layer of public accountability for individual journalists. A low score might follow a reporter across jobs, while a high score could become a bargaining chip in negotiations. The criteria the AI uses - source attribution, tone, right of reply - are not neutral; they embed assumptions about what good journalism looks like. Writers who work in opinion, investigative, or beat reporting may find their work assessed differently by a model trained on a particular view of rigor. Understanding the mechanics behind such scoring, and the interests of those who build it, will matter for anyone whose livelihood depends on bylines.
Your membership also unlocks:
