Anthropic’s Dario Amodei recently urged a coordinated pause on frontier AI development, citing existential safety risks. He asked for antitrust exemptions to enforce this slowdown, a move backed by OpenAI’s Sam Altman and Elon Musk. But the industry isn’t buying the altruism. Cohere CEO Aidan Gomez fired back in a scathing blog post, labeling the proposal “a cartel by another name.” He argues that granting a handful of labs regulatory immunity creates insurmountable barriers to entry. Only firms with massive compute resources and deep government ties could comply, effectively locking out smaller competitors.
Gomez points out a glaring conflict of interest: the “independent” evaluators proposed to audit these systems are funded by the very companies they monitor. By defining risk strictly as a function of scale, the framework ensures only the biggest players get to define what is safe. While these giants debate hypothetical superintelligence, Gomez notes that cheap voice cloning tools are already draining pensioners’ accounts. The focus on distant existential threats feels like a smokescreen for immediate market consolidation.
The pushback extends beyond business rivals. Niels Rogge, an engineer at Hugging Face, dismissed Amodei’s concerns as “bizarre nonsense.” He suspects the real target is open-weight models from competitors like China’s DeepSeek, which threaten the closed-source monopoly. Rogge highlighted that open-source communities actually helped defend against hacks from closed-model autonomous agents. Meanwhile, White House AI czar David Sacks called the duopoly’s motives suspect. He suggested the slowdown is less about safety and more about limiting product liability after recent security breaches.
Political reactions further fracture the consensus. President Donald Trump labeled AI takeover warnings a “hoax,” comparing them to climate change scams. He claimed China would not hinder AI progress, dismissing US caution as political theater. Beijing has indeed rejected the slowdown proposals, viewing them as a tactic to cement American dominance. With major global players refusing to participate, any unilateral slowdown may only handicap US innovation while leaving geopolitical rivals free to accelerate. The safety argument is crumbling under the weight of obvious competitive maneuvering.
