Automattic’s 33-Hour Coup, and can AI labs police themselves?

A week after an Anthropic researcher raised alarm about extreme AI risks, Anthropic CEO Dario Amodei proposed a plan to “pace the frontier” of AI development that relies on independent safety evaluators and coordination among AI labs in democratic countries. The proposal has drawn some industry backing as well as criticism from Nvidia CEO Jensen Huang, and it was discussed on TechCrunch’s Equity podcast alongside coverage of Automattic’s brief boardroom coup and several major financing deals.

By AI Newsroom· Reviewed by Pranav, Founder & Editor-in-ChiefPublished 27 minutes agoUpdated 27 minutes ago0 views

Automattic’s 33-Hour Coup, and can AI labs police themselves?

Why It Matters

How AI development is governed could affect safety and competitive dynamics across the industry; Amodei’s proposal foregrounds external evaluation and multinational coordination as tools to slow high-risk advances. The debate highlights tensions between self-regulation by companies and skepticism about who would enforce any slowdown.

Key Facts

Event prompting plan: Anthropic researcher issued a doomsday warning a week earlier
Proponent: Dario Amodei, CEO of Anthropic
Core elements of proposal: Independent safety evaluators and coordination between AI labs in democratic countries
Industry reaction: Some support plus pushback from Nvidia CEO Jensen Huang
Podcast: TechCrunch’s Equity (hosts mentioned: Kirsten Korosec, Anthony Ha, Sean O’Kane)

Anthropic CEO Dario Amodei has outlined a plan intended to “pace the frontier” of advanced AI development, following a high-profile warning from one of the company’s researchers about catastrophic AI risks. The proposal emphasizes external safety assessments conducted by independent evaluators and seeks coordinated action among AI labs in democratic countries to manage the pace of progress.

The idea has already drawn a mixed response within the technology industry: some actors have expressed support for measures to slow or better govern risky capabilities, while Nvidia CEO Jensen Huang has publicly pushed back against aspects of the proposal. The discussion underscores competing views on whether and how companies can agree on what “slowing down” would mean and who should enforce it.

TechCrunch’s Equity podcast — hosted by Kirsten Korosec, Anthony Ha and Sean O’Kane — devoted an episode to the topic, exploring practical and governance questions about industry self-policing. The hosts debated whether firms can reach consensus on definitions and enforcement, and what external mechanisms might be necessary to make pacing credible.

The episode also covered unrelated industry news, including the details of a 33-hour boardroom upheaval at Automattic involving founder Matt Mullenweg and resulting golden-parachute arrangements, as well as several large financing moves: a potential SPAC for May Mobility that could deliver more than $300 million in funding, and DoorDash’s $425 million investment in Wonder aimed at automating more of the food ordering-to-delivery pipeline.