WIRED's Kylie Robison on the OpenAI AGI paper suppression, the $100M offer debate, and what comes next in the talent war

Summary

The weekend of June 28–30 produced a rapid sequence of AI talent war disclosures, with Wired's Kylie Robison at the center of the reporting. The first scoop, published Friday, revealed that an internal OpenAI research paper defining AGI has been suppressed, with sources telling Robison that ongoing Microsoft deal negotiations were cited internally as the reason for withholding it. The Wall Street Journal and Reuters have separately reported on the tension surrounding those negotiations.

Four OpenAI researchers departed for Meta's new lab on Saturday. On Sunday, OpenAI chief research officer Mark Chen circulated an internal memo characterizing Meta's recruiting push as essentially a home invasion. Monday morning, Mark Zuckerberg published a public list of new hires, making the poaching campaign fully visible.

The $100M Offer Question

The figure that has dominated coverage — $100 million per researcher — carries real uncertainty beneath the headline. Both the Wall Street Journal and the New York Times printed the number, suggesting their sourcing clears those outlets' editorial bars, but Robison says she has not personally seen an offer letter with that figure and would require one before reporting it herself.

The more relevant framing, as Robison notes, is structural. Compensation packages of this kind are almost certainly multi-year earnouts with clawback provisions, not lump-sum transfers. The Information has reported on the underlying researcher salary landscape at OpenAI, and researchers who joined when the company was valued at $40–80 billion are already sitting on significant paper gains, reducing the marginal pull of Meta's offers. Framed differently, spending $5 billion to assemble 50 researchers critical to a multi-trillion-dollar company's core product is not extraordinary by precedent — Meta paid roughly $380 million per employee when it acquired WhatsApp's ~50-person team for $19 billion in 2014.

The Memo's Strategic Weakness

Chen's internal memo has drawn criticism for its tone. Rather than leaning on OpenAI's demonstrated resilience — the company retained dominant consumer market position despite losing Daario (Andrej Karpathy), Elon Musk (who founded xAI), and the founders of Thinking Machines and SSI — the memo focused narrowly on the mechanics of how Meta's offers are structured and urged employees to surface any outreach. Robison agrees the defensive, operational framing read as reactive rather than rallying, likely reflecting the pressure of Zuckerberg personally calling researchers and the timing of an upcoming company-wide break.

What Comes Next

Meta still has approximately 50 additional hires it is targeting for the new lab. Robison expresses skepticism that Zuckerberg will land researchers of Ilya Sutskever's caliber. Her argument is that the most mission-driven AGI researchers — those who genuinely believe they are building transformative or existential technology — may not find Meta's value proposition compelling regardless of compensation. Building superintelligence for Zuckerberg is a different motivational frame than building it under a safety-first or independent research mandate.

Two names remain unresolved. Daniel Gross, reported by The Information as a potential hire alongside Nat Friedman, did not appear on Zuckerberg's published list, and Robison could not confirm his status. A team based in Switzerland was also absent, which Robison attributes speculatively to EU employment law complications rather than a deal falling through.