Anthropic CEO Dario Amodei reiterated his call for a slowdown in AI development at Salesforce's Dreamforce convention in San Francisco on Tuesday. He used an automotive analogy, suggesting that safety incidents should prompt all companies to review their own practices.
Amodei proposed three actions: embedding third-party evaluators within AI companies, coordinating safety standards among democratic countries, and eventually achieving larger global coordination. He stated that Anthropic has committed to independent evaluators and plans to discuss the other steps with the industry.
Shortly after, Nvidia CEO Jensen Huang argued against slowing down AI development, stating, "Run as fast as you can." Huang advised AI companies to ensure products are safe before release rather than seeking government intervention. He also dismissed concerns about AI destroying jobs as "completely nonsense."
The calls for a slowdown follow warnings from Jacob Coxon, a former Anthropic researcher, who publicly resigned over concerns that AI could lead to human extinction within a decade. This has sparked debate in Silicon Valley and gained regulatory attention.
Amid these discussions, both OpenAI and Anthropic are preparing for initial public offerings. OpenAI's Sam Altman reportedly plans to delay his company's IPO until next year due to safety concerns.