The trick: Rented Halo

Claude moved a bound that had not moved in years: 41.6 to 67.2.

Journal reviews of the paper so far: zero.

Issue 713 August 20265 receipts3 min

Anthropic says an unreleased research Claude improved the longstanding lower bound for the fraction of Riemann zeta zeros that satisfy the hypothesis from 41.6% to 67.2%, burning 650 failed ideas and 31 million output tokens across two agentic sessions.

Before you read on. Your call?

the evidence is unusually strong for a same-day lab announcement. Two Anthropic mathematicians validated the paper, Claude produced a formally verifiable proof, and outside experts Brian Conrey and Dan Goldston examined it favorably. But the paper has not been submitted to a journal, a short-notice examination is not peer review, and the whole run happened on an unidentified model nobody can rerun. The theorem is checkable. The capability story around it is not.

67.2%THE NEW BOUND

650IDEAS THAT FAILED BEFORE ONE WORKED

0JOURNAL PEER REVIEWS OF THE PAPER SO FAR

There’s more to this story.

Membership opens the full investigation, the strongest counterargument and what to do with what you’ve learned.

Start your free month →

First membership: 30 days free, then A$89 a year. One introductory trial per customer. Card required; renews annually until cancelled. Cancel before the trial ends to avoid the first charge. Already a member? Sign in

The trick has a name

We call it Rented Halo: the achievement is real, the drama around it is borrowed. You'll see it again. Learn to spot it →

Say this in tomorrow's meeting“The 67.2 is probably real math. The two-session autonomy story comes from a model nobody outside Anthropic can run.”

Receipts

Supports anthropic.com: it has increased this bound from 41.6% to 67.2%. Two mathematicians at Anthropic studied and validated Claude's paper
Context anthropic.com: Claude also produced a formally verifiable proof of its result. We are grateful to Brian Conrey and Dan Goldston, two experts in this area, who generously examined the paper on short notice. We don't expect that the techniques Claude used will lead to proving the Riemann hypothesis.
Context anthropic.com: Claude generated and tried 650 ideas, none of which worked
Refutes kingy.ai: It is not a proof of the Riemann hypothesis, has not yet passed conventional peer review, and cannot be reproduced end to end because Anthropic used an unidentified research model.
Refutes xenospectrum.com: This result has not been submitted to a journal and has not undergone peer review. The review by Conrey and Goldston is described as a "favorable examination on short notice," which is different in nature from a formal peer review.

Open the Receipts Pack → What each source proves, every figure traced, and what would change our verdict.