Elon Musk backs Dario Amodei's arguments about pacing the frontier, saying “Dario is right”
Anthropic CEO Dario Amodei on Saturday called for the immediate slowdown of artificial intelligence development, warning that recent advances in AI have begun to outpace the industry's ability …
PoliticoBen Johansen
Context & Ripple Effects
Amodei has framed pacing as a way to preserve continued training while allowing safeguards and outside checks to catch up, rather than as a halt to progress. Anthropic has paired that position with a commitment to permanent access for third-party evaluators, extending its earlier call for federal transparency standards over a broad regulatory moratorium.
Musk’s endorsement gives Amodei’s safety case support from a prominent figure outside Anthropic, an unusual convergence given the competitive framing that has often surrounded frontier-model development.
First-order effects
Amodei’s proposal gains a higher-profile public validator, strengthening Anthropic’s case for development schedules tied to safeguards and independent evaluation.
Anthropic’s evaluator-access commitment becomes a more concrete reference point in the pacing debate than a general call to slow AI development.
Second-order effects
Frontier-model developers face greater pressure to explain how they test safeguards and whether external evaluators can verify their claims, not merely to state voluntary principles.
If rival labs adopt comparable access and verification practices, frontier AI competition shifts partly from model-release speed toward demonstrable governance capacity.
The debate points toward frontier AI institutionalization: progress remains permitted, but increasingly conditioned on auditable safety processes and coordination among companies and governments.
The trend: Frontier AI governance is moving from abstract slowdown demands toward operational commitments that link deployment pace to independent verification.
I think one reason some are skeptical of this wave of Do Something-ism is less about the foundational claims (I think we can all agree recursive models are dangerous regardless of whether they are “AGI
Elon, perhaps the last person I would've thought would agreed that the development of frontier models should slow, has said he agrees with Dario's “We Must Pace The Frontier” blog post Nominally, that means OpenAI, Anthropic and xAI have now signalled their willingness to slow. W…
This is a big update for me personally. Not that he agrees, but that he supports pacing. He never signed the letter. This does change my model of things a little.
Strong disagree. If the US labs slow down, either voluntarily or due to regulation, it won't be long before other countries with relentless ambition take the lead in AI. …
I think there's actually quite an easy and obvious way to solve the collective action problem of how to make the AI companies stop conducting risky research, and it's to make it very clear that they're legally liable for anything their models do. www.politico.com/news/2026/09...