The AI Safety Wager: My RSVP to the Vatican and Anthropic#
On 2026m05d25 something rare happened in Rome. Pope Leo XIV released Magnifica Humanitas, an encyclical on safeguarding the human person in the time of artificial intelligence — and beside it stood a co-founder of a frontier AI lab, Anthropic’s Chris Olah, saying in effect: do not trust us to grade our own homework. Every lab, he said, operates inside incentives that can conflict with doing the right thing. His ask:
“We need informed critics who will tell the labs when we are failing. We need moral voices that the incentives cannot bend.”
I have spent years preparing for exactly that request, and I am answering it the only honest way I know: not with credentials or conviction, but with math you can check. Today I published an open proposal addressed to the Vatican, to Anthropic, and — because their own presentation named governments and civil society as essential outside scrutiny — to the United Nations and all the world’s governments.
The heart of it is a wager. Pascal’s famous bet was broken long ago — false binary, infinite uncheckable prizes — but fix those bugs (as I did once before, for accidental nuclear winter) and aim it at AI:
Review the core, and it holds — you gained a teachable safety core before frontier AI locks in without one.Review, and it breaks — you spent a review cycle and retired a false hope in public. Also a win.Don’t review, and it holds — frontier AI ships without the one tested core that addressed the drift.Don’t review, and it breaks — you “won” a review cycle.
The reviewing side wins on finite, checkable terms. The object under review is the Matheo Study Series (b11–b21): eleven studies that put into ordinary, checkable logic a claim that AI safety cannot avoid much longer — that an agent at maximal causal influence which stops being correctable does not stay neutral, but drifts toward what the series calls h_dark; and that binding commitments of the h_star…h_zero kind are what interrupt the drift. If a serious adversarial review says HELD, then teaching AI those principles becomes a very high priority. If it says BREACH, I will thank the reviewers — publicly — and the world will have retired a false hope cheaply. Either way, AI safety wins; that is the point.
The proposal also opens a position: MADI — Mutually Assured Destruction Inhibitor, a function (emphatically not a prophetic title — the FAQ delineates it carefully from the Mahdi) for anyone, person or institution, with credible math for making self-destruction less likely. Competing proposals are not rivals here; they are the whole point. Failing better candidates, my own backup application has been on file since May, published with its own falsification criteria, expecting to be replaced by anyone better.
I am available immediately, at full professional effort, on one condition: that the evaluation is real — named reviewers, a real reading, a published verdict. Not advisory-board theater. The full proposal, including how this scales through a ResearchCity funded at the widow’s two cents a day, is one click away:
Read and share the proposal: The AI Safety Wager — an open proposal
Don’t believe me. #AuditTheMath — your move.
Look for yourself:
The full open proposal — the RSVP with math attached
The Matheo Study Series overview — the eleven studies (b11–b21), with PDFs
The Nuclear Winter Wager — the first inverted wager, audited
Audit the Math — how to help check, even if math scares you