One essay, a hundred replies

Dario Amodei published "We Must Pace the Frontier" on a Saturday morning. Within days, rival CEOs, the president, senators and former cabinet officials had all weighed in.

Amodei's three steps

He proposed embedded third-party evaluators inside AI companies, coordination on standards among frontier labs in democracies, and talks with authoritarian governments "to the extent this is possible".

Pacing is not stopping

Amodei drew the line early: slowing down means giving companies enough time to align and safeguard models, and letting outside evaluators confirm it. Training continues.

“does not mean halting model training or technical progress”

Anthropic goes first

Rather than wait for anyone to agree, Amodei said Anthropic is unilaterally committing to step one, the embedded third-party evaluators.

Altman says yes

OpenAI's CEO agreed within hours, called independent evaluators a great idea, and said pacing had already been a primary topic internally in recent weeks.

“I agree with Dario that we need to pace the frontier”

OpenAI asks for federal rules

By Sunday Altman wanted a federal framework with consistent safety requirements, plus shared standards on misalignment and monitoring that other labs could build on.

Altman's two failure modes

He named the ways this goes very badly: losing control of the future to AI, or too much concentration of power. His answer was a narrow middle path between them.

“walking a narrow middle path”

Hassabis and Musk sign on

Demis Hassabis said the direction is correct and pointed back to his own July proposal for an AI standards body. Elon Musk reposted the essay with three words: Dario is right.

“Dario's essay points towards the right path forward”

Trump rejects the premise

In an all-caps Truth Social post, the president said the only guardrail AI needs is a smart president, accused Amodei of playing a perfect little angel, and called the slowdown push a conspiracy that only China benefits from.

“WHOEVER WINS AI, WINS!”

The China argument takes over

Vance called labs begging for regulation a bit of a trojan horse. House Speaker Mike Johnson ruled out a moratorium because China would overtake the US, though he offered to convene the labs tomorrow.

“We cannot put a moratorium on this because China will overlap us”

Democrats want hard limits

Bernie Sanders backed a column titled "Pause AI for Humanity's Sake". Kamala Harris called a slowdown critical to the public interest. Pete Buttigieg wants rules, limits, and a kill switch for rogue AI.

“There need to be safeguards, there need to be rules, there need to be limits”

The skeptics on both flanks

David Sacks told the labs to slow down on their own and stop pretending the motive is altruism. Lina Khan noted there is no AI exemption from existing product liability law. Yann LeCun just said everyone should make fun of them.

“stop pretending you need anyone else's permission”