'If You Mean It, Open the Weights': Viral Open Letter Challenges Dario Amodei's AI Slowdown Call

A viral open letter challenges Anthropic CEO Dario Amodei to back his AI slowdown call with a radical proposal: open weights for every public model.

Hours after Anthropic CEO Dario Amodei published his "We Must Pace the Frontier" essay calling on governments to slow the AI race, a pointed response began climbing Hacker News, eventually becoming one of the most upvoted stories of the day. Its argument, laid out in an open letter by writer and developer Jacob Gold, is disarmingly simple: if Amodei is sincere about slowing AI down, there is one law he could call for that would actually do it — and no other lab executive could propose it. Require every AI model a company offers to the public to be released as open weights.

"I believe you're being sincere," the letter opens, addressing Amodei directly, "so I'm asking you to call for a different law, one that would actually slow things down, even though it would require real sacrifice on your part." The letter has struck a nerve precisely because it engages with the strongest version of Amodei's argument rather than dismissing it — and because it lands in the middle of the most consequential policy debate in AI today, a debate we follow daily in our AI policy coverage.

What Amodei proposed, and why Gold says it won't work

Amodei's essay committed Anthropic to embedded third-party evaluators and asked governments to require every other frontier company to match that commitment. Sam Altman agreed within hours — "I agree with Dario that we need to pace the frontier," he said, according to Politico, with OpenAI matching the evaluator commitment as well. On its face, that is an unprecedented alignment: the heads of the two most valuable AI labs in the world publicly asking to be slowed down.

Gold's letter argues it is precisely this consensus that should make observers suspicious. "Every regulation you've asked for ends in capture," it warns. Embedded evaluators, compute thresholds, industry coordination with antitrust waivers — "All of it gets written with the help of the current frontier labs, because nobody else understands the technical details well enough to write it." Once such rules become law, every incident adds a rule and none ever gets removed, and the giants can afford the compliance teams that keep smaller challengers out. "Instead of slowing you down," the letter tells Amodei, "it protects your position and profits."

The open-weights counterproposal

The alternative Gold sketches is deliberately blunt. Any AI model a company offers to the public would have to ship with its weights published. Internal and research models would be untouched, and governments would retain access to unreleased systems — but Anthropic would have to publish the weights of its next frontier model the day it was released, and OpenAI would have to do the same. "There's nothing in it to game or dodge," the letter concludes, "even for you."

The mechanism is financial rather than bureaucratic. As Gold explains, funding for frontier model development depends on valuations that assume the weights remain proprietary. A mandate flips that assumption: if every public release is open, competitive advantage flows to whoever trains the best model rather than whoever sells access to it, the premium embedded in private valuations deflates, and the money available for the next training run shrinks — "at every lab at once, without a regulator deciding anything." It is, in effect, an argument that the market's expectation of closed models is the engine of the race, and that removing the expectation removes the engine.

Why Amodei, specifically

The letter's case for why Amodei is the one person who could make this proposal is built on his record, and it is notably respectful. Gold credits him with leaving OpenAI at the end of 2020 to found Anthropic "when the safe career move was to stay," with advocating regulation of his own industry for years, and with pushing export controls that shrink his own market. He notes that Anthropic structured itself as a Public Benefit Corporation with a legal mission — "the responsible development and maintenance of advanced AI for the long-term benefit of humanity" — that would allow its leadership to absorb the financial hit such a law would deliver. Amodei's own essay, Gold observes, says Anthropic advocates for regulation "even when this gets us accused of hype, 'doomerism', or regulatory capture."

Skeptics will note the obvious counterarguments, and the letter does not fully answer them: open weights make misuse easier as well as progress harder, an open-weights mandate would face fierce opposition from every lab's investors, and critics of Amodei's earlier positions on Chinese open models will hear echoes of old fights in the proposal. But the letter's core challenge does not depend on its policy being adopted. It depends on a tension it names explicitly: a lab that asks governments to slow everyone else down, while its own valuation depends on staying ahead, has not yet asked for anything that costs it anything.

A debate that is now unavoidable

Whether Amodei responds is almost beside the point. Within a single day, the frontier labs moved from competing on capability to agreeing on the need for pacing, and the public response has moved from asking whether AI is dangerous to asking whether the people warning about it can be believed. Gold's letter — a few hundred words of policy detail wrapped around one uncomfortable question — has become the sharpest expression of that skepticism, and its virality suggests the pacing debate will be fought not just in Washington but in public, on platforms like Hacker News, where the open-weights community has been making variations of this argument for years.