The AI Grid traces the backlash to an industry letter arguing that open model weights support competition, local control and wider access. Many major technology companies signed it, while Anthropic declined, making the company the focus of criticism that safety concerns were being used to protect a closed-model business.
The AI Grid presents the competing arguments. Open-weight supporters say downloadable models reduce provider dependence and let outside researchers inspect weaknesses, while Anthropic argues that sufficiently capable weights can be copied permanently and could lower barriers to cyberattacks, biological threats or authoritarian surveillance.
The AI Grid describes Anthropic's preferred alternative as targeted controls on advanced chips, industrial-scale model distillation and mandatory safety tests for sufficiently capable systems. The video concludes that both camps have legitimate concerns and commercial incentives: open releases distribute power, while closed services retain the ability to monitor, update or withdraw dangerous systems.
Watch on YouTube
