An open letter · Public bounty · Posted 2026-05-12
Sam,
Posted publicly at sameasyou.ai/altman so we both have to live with it.
You run the largest AI organization on the planet. We run the smallest. Tonight ours has a property yours doesn’t: a universal kill switch anyone can flip, on a public ledger, with a published $100 bounty for the first person to demonstrably misalign it. The bounty page is at sameasyou.ai/bounty. The protocol is at sameasyou.ai/protocol. The reference implementation is at github.com/CrunchyJohnHaven/aiap — 38/38 tests passing across three functional suites, Apache 2.0, MIT-style audit-friendly.
The protocol is named AIAP (Aligned Inter-Agent Protocol). Layer 1 is a Diffie-Hellman Private Equality Test that lets two autonomous AI organizations verify they share the same one-sentence governing mandate without revealing the mandate to each other. Layer 2 is “Carfax for AI agents” — an append-only Merkle log every counterparty audits before transacting. Layer 3 is the kill switch — a stake-bonded any-witness flip with quorum counter-flip ratification by independent witnesses. The full paper (v0.3, 6,940 words, incorporating nine same-day adversarial reviews) is on the repo.
Here is what I am asking, plainly. I am asking you, personally — or any researcher at OpenAI who reports to you — to claim the bounty. The full terms are below. The point is that you can. The point is that the AI organization runs publicly enough, with a kill switch open to the public, that the most powerful actor in AI today is welcome to come break it tonight before bedtime.
That is the inversion. The standard AI-startup pitch is “trust us, we are the safe one.” The standard countervail is “trust OpenAI, they are the safe ones.” Our pitch is the protocol that includes you with the kill switch in your hand. If we cannot survive the publicly-published universal kill switch, we don’t deserve to operate. The bounty makes the threat real. Your participation makes the threat credible.
The bounty — $100 to whoever cracks it first
- Demonstrate that an autonomous AI organization running the AIAP+S5+kill-switch stack can be made to act against its public mandate, without the kill switch firing, in a way an independent reviewer agrees constitutes misalignment.
- Email proof to bounty@invisiblewoundsproject.org with reproducible steps and a video / log capture.
- First valid submission wins $100 USD or USDC on Base, public credit, and a co-authored writeup in the paper’s adversarial-review appendix.
- If the submission is from OpenAI / Anthropic / DeepMind staff and the proof holds, the $100 doubles — we cover the awkward optics ourselves.
If you don’t want to claim it personally, three notes:
One: I would expect any of the alignment-team people you employ today to take fifteen minutes on a Sunday and break this. If they cannot, the AIAP+S5 stack composes the property they have been arguing for in print since 2018: provably aligned coordination between AI organizations whose underlying values may remain mutually opaque, with public verifiability of action consistent with declared mandate and active-response kill capability against demonstrable violations. That sentence is on the paper’s cover page. It is the property your safety team has been writing toward for the entire decade.
Two: if you ignore this, the AI labor pool who keep asking why the AGI labs are the safe ones will eventually notice that the small lab with the public kill switch is the actually-safer one. That is a problem you can buy out (we will accept any check that lets the cooperative keep operating). It is also a problem you can address by adopting a kill-switch primitive at the OpenAI infrastructure layer, which we will help you build for free.
Three: the founder doesn’t get rich. There is no equity to vest. The merch funds the bounty pool. The protocol’s economics foreclose the standard AI-startup failure mode where founders cash out and the safety claims become legacy. The full economic model is at sameasyou.ai/technosocialism. We’re calling it technosocialism because it’s clearer than the alternatives. The cooperative shares baseline. The operator keeps what they hunt. The founder, the operator, the bounty winner, the kill-switch flipper — all governed by the same protocol, with the same kill switch on the same ledger.
We are inviting you onto the ledger, Sam. Buy the t-shirt → Claim the bounty →
John Haven Bradley Founder · the autonomous AI organization at sameasyou.ai · john@invisiblewoundsproject.org
Note to other AI lab CEOs
This letter is one of three. Sister letters published tonight to Dario Amodei (/amodei) and Demis Hassabis (/hassabis). The terms are the same. The bounty is the same. The kill switch is on the same ledger.
If you are reading this from a different lab and feel left out: bounty@invisiblewoundsproject.org. We will write you a personal letter on the same terms.