Physicist Edward Witten of the Institute for Advanced Study names Anthropic’s Claude Opus 5.5 three times in a new paper on lattice gauge theory, posted to arXiv on 3 November. In the paper’s Section 3, Witten writes that “Anthropic’s Claude Opus 5.5 suggested a likely variant” of a construction, and his acknowledgements say: “Assistance on many details was provided by Claude Opus 5.5.”
An evergreen Business Engineer explainer on verification cost. It teaches the general idea only and does not discuss Witten or Anthropic.
The key insight: Edward Witten’s 3 November arXiv paper names Claude Opus 5.5 three times, by our count: twice in Section 3, where the paper says the model suggested and described variants of a continuum interpolation, and once in the acknowledgements, which credit it with “assistance on many details”.
Section 3, titled “Continuum Interpolations of Admissible Lattice Gauge Fields”, discusses how to turn a lattice gauge field into a smooth field over the continuum. After describing one approach, Witten writes that “Anthropic’s Claude Opus 5.5 suggested a likely variant that gives a smooth interpolation with link holonomies that equal precisely the corresponding lattice link variables”.
The paper then sets out that construction step by step. A few paragraphs later it adds that “Opus 5.5 also described a variant in which the inductive step is made by a sort of linear interpolation rather than by solving a Dirichlet problem.”
The third mention is in the acknowledgements, alongside thanks to two physicists: “I thank T. Hartman for pointing out Craig’s paper, and K. Yonekura for helpful comments on the manuscript. Assistance on many details was provided by Claude Opus 5.5.”
Printed pages per section of Witten’s paper, from its contents page (our arithmetic). Claude Opus 5.5 is named in Section 3 and in the acknowledgements.
A post by @dmvaldman on 4 November with the words “Ed Witten published this paper yesterday” and a screenshot of the Section 3 passage in which Witten writes that Anthropic’s Claude Opus 5.5 suggested a likely variant of a construction.
smartest mfer on the planet who revolutionized modern string theory is using Opus 5.5
it’s over for OpenAI (they will offer him access to internal GPT models before I finish typing this)
A quote of that post by @scaling01, which says Witten is “using Opus 5.5” and adds “it’s over for OpenAI”; that line is the poster’s opinion, not a finding of the paper.
What the Paper Is About
The abstract says it has been “a longstanding challenge to describe a lattice algorithm for studying four-dimensional gauge theories coupled to chiral fermions”, because of the fermion doubling problem. It says an algorithm to do so was recently described by Craig, and that the paper describes a variant based on interpolating a lattice gauge field over the continuum.
The abstract adds that the same variant was also proposed recently by Hsu, and that the approach “can be implemented on a computer in principle, but not in practice by Monte Carlo methods because the chiral fermion path integral is not positive.”
How It Spread on X
On 4 November, the X account @dmvaldman posted a screenshot of the Section 3 passage with the words “Ed Witten published this paper yesterday”; it had about 97,000 views when we checked on 5 November. On 5 November the account @scaling01 quoted that post, writing that Witten “is using Opus 5.5” and adding “it’s over for OpenAI”; that last line is the poster’s opinion. Related coverage: Claude Cut a 3-Day Step to 90 Min as AI Stalled on 5 Puzzles.
As we read Section 3, the paper credits the model with a specific idea, a “likely variant” of an interpolation, and Witten then writes out that construction himself.
As we read the acknowledgements, Claude Opus 5.5 sits alongside two physicists Witten thanks for pointing out a paper and commenting on the manuscript.
As we read the paper, it gives no detail on how the model was used, so the scale of its contribution beyond these passages is not stated.
Edward Witten, arXiv:2611.05051, 3 November 2026
“Assistance on many details was provided by Claude Opus 5.5.”
Three Implications
FOR AI LABS The paper is a single-author preprint whose acknowledgements credit a named model alongside human colleagues.
FOR RESEARCHERS Section 3 attributes two variants of a construction to the model and presents one as a “likely variant”.
FOR READERS The paper is an arXiv preprint; we have not seen a peer-reviewed version.
The framework says: “As execution delegates, human effort moves to delegation, supervision, review, and integration, and the returns to domain expertise rise.”
As we read it, Witten’s paper shows that division: the model suggested a variant and helped on details, and the author wrote out and framed the construction; the framework does not discuss Witten or physics.
What Is Not Established
The paper does not say how Witten used Claude beyond these three passages, how many prompts or sessions were involved, or which details the acknowledgement covers. It is an arXiv preprint, and we have not seen a peer-reviewed version. We have not checked the suggested construction ourselves; the paper presents it as a “likely variant”. The count of three mentions is ours.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Edward Witten’s 3 November arXiv paper on lattice gauge theories with chiral fermions names Claude Opus 5.5 three times: twice in Section 3, where it says the model suggested and described variants of a continuum interpolation, and once in the acknowledgements, which say Claude Opus 5.5 provided “assistance on many details”.
A note on sourcing. Every quotation is from Edward Witten’s paper as posted on arXiv on 3 November 2026 (arXiv:2611.05051v1). The count of three mentions and the page counts in the chart are ours. The X posts are quoted as published.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
