In a critique of current AI industry trends, David Sacks challenges the calls from OpenAI and Anthropic leaders to "pace the frontier" via external regulation. Sacks argues that these two companies currently maintain a duopoly on frontier intelligence, measured by market share, revenue, and model capability, driven by recursive self-improvement. He contends that these firms do not require government permission or the suspension of antitrust laws to manage their development cycles. Sacks specifically questions the independence of evaluators like METR, noting their ties to Anthropic’s staff and investors, and warns against using regulatory approval processes to supersede existing product liability laws. He suggests that the push for "alignment" and slower deployment may be driven by a desire to mitigate massive product-liability exposure from cyberattacks or unpredictable model behavior, rather than purely altruistic motives. Sacks posits that slowing down could be a strategic business move to increase reliability and predictability for customers, similar to the lessons learned following the Hugging Face incident. Ultimately, he warns that demanding specific regulatory frameworks as a condition for slowing development resembles regulatory capture or political maneuvering, particularly given the unlikely participation of China in any global agreements. He suggests that if these companies choose to slow down, they should do so voluntarily to build public goodwill rather than forcing a legislative mandate.