Anatoly Yakovenko, co-founder of Solana, criticized a proposed framework to slow AI development this week, claiming the movement prioritizes corporate market consolidation over genuine safety considerations.
Yakovenko responded on X to public endorsements of the slowdown proposal from Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and Elon Musk. Yakovenko posted "Profitability at $1 trillion mcap," suggesting the deceleration effort centers on protecting major AI companies as they approach trillion-dollar valuations rather than addressing legitimate safety issues.
Amodei's Proposed Framework
Amodei outlined his position in an essay titled "We Must Pace the Frontier," advocating for decelerating AI capability expansion while continuing research efforts.
The framework includes three components: granting external evaluators access comparable to internal safety teams, establishing unified safety standards among leading AI firms in democratic nations, and pursuing global cooperation including potential arrangements with China regarding AI-enhanced weaponry and automated model enhancement processes.
Amodei acknowledged potential antitrust concerns and suggested limited government authorization might be necessary to facilitate industry coordination.
On September 12, Altman expressed support, stating "I agree with Dario that we need to pace the frontier." He committed to incorporating independent evaluators with full employee access and indicated OpenAI would provide additional details.
Musk offered brief endorsement, sharing Amodei's essay with the comment "Dario is right," without elaborating on specific aspects of the proposal.
Industry Opposition
David Sacks, a former White House AI and crypto adviser, argued that Anthropic and OpenAI can voluntarily reduce development pace without imposing regulatory restrictions on the broader AI industry.
Sacks characterized the two firms as holding a "duopoly on frontier intelligence" and warned that centralized regulation could harm smaller startups and open-source projects that are narrowing performance gaps with established leaders. He also cautioned that international AI agreements face low success probability given China's likely non-participation, and suggested constraining American AI laboratories could constitute voluntary technological disadvantage.
Neither Anthropic nor OpenAI has issued direct responses to the objections raised by Yakovenko or Sacks.
Financial markets showed no significant reaction to the debate. Industry analysts noted that reducing frontier model development speed does not necessarily correlate with decreased overall AI infrastructure investment.
As of September 13, Anthropic has pledged to implement external evaluation procedures but has not announced an implementation timeline or identified specific evaluators.
