Key Highlights
- Anatoly Yakovenko, co-founder of Solana, claims the AI development slowdown initiative centers on “profitability at $1 trillion mcap”
- Dario Amodei, CEO of Anthropic, released an essay advocating for reduced AI advancement speed due to safety concerns
- Both Sam Altman of OpenAI and Elon Musk endorsed Amodei’s slowdown framework
- David Sacks, ex-White House AI policy adviser, argued that AI companies should self-regulate without imposing industry-wide restrictions
- Financial markets remained stable despite the slowdown discourse at Monday’s trading session
Solana co-founder Anatoly Yakovenko launched a sharp critique this week targeting leading AI executives after Anthropic’s Dario Amodei, OpenAI’s Sam Altman, and Elon Musk publicly endorsed reducing the pace of AI advancement.
Yakovenko’s rebuttal was notably concise. On X, he posted just four words: “Profitability at $1 trillion mcap.” The statement referenced no particular organization nor included supporting financial evidence.
The implication was unmistakable. Yakovenko appeared to argue that calls for AI deceleration stem not from genuine safety considerations but from efforts to cement dominant market positions once companies achieve trillion-dollar market capitalizations.
In a subsequent post, Yakovenko mentioned instructing his Codex to reduce token usage, seemingly satirizing the concept of artificially pausing AI advancement.
Amodei’s Framework Explained
Amodei released an essay entitled “We Must Pace the Frontier,” advocating for decelerating AI capability advancement rather than halting research entirely.
The framework consists of three phases. Initially, Anthropic would grant external assessment organizations continuous access comparable to internal safety personnel. He cited METR, the Model Evaluation and Threat Research organization, as a potential partner.
The second phase involves coordination among frontier AI firms in democratic nations on unified safety standards. Amodei recognized potential antitrust implications and suggested limited governmental authorization might be necessary.
The final phase encompasses global cooperation, potentially including arrangements with China, spanning restrictions on AI-powered weaponry to regulations on automated model enhancement.
On September 12, Altman publicly supported the framework, stating “I agree with Dario that we need to pace the frontier.” He committed to incorporating independent assessors with employee-equivalent access and promised forthcoming details from OpenAI.
Musk’s endorsement was succinct. According to Reuters, he reshared Amodei’s essay with the comment “Dario is right,” without elaborating on which elements he backed.
Sacks Questions Regulatory Approach
David Sacks, previously the White House AI and cryptocurrency policy adviser, contended that Anthropic and OpenAI could independently reduce their development pace without subjecting the broader industry to regulatory frameworks.
He characterized the pair as forming a “duopoly on frontier intelligence” and cautioned that top-down regulations might devastate emerging startups and open-source initiatives that have been narrowing the technology gap with established laboratories.
Sacks further questioned the viability of international AI agreements, asserting that China would not honor such commitments. He cautioned that constraining American research facilities could constitute voluntary technological disadvantage.
Neither Anthropic nor OpenAI has issued direct responses to the criticisms from Yakovenko or Sacks.
Financial markets displayed no significant turbulence. Industry analysts observed that decelerating frontier model development doesn’t necessarily translate to reduced overall AI infrastructure investment.
As of September 13, Anthropic has pledged to implement external evaluations but hasn’t announced a timeline or designated an assessment organization for the initiative.


