The open-weights argument becomes the year's real fight about intelligence

On 24 July an industry statement backing open-weight models drew signatures from Microsoft, Google, Meta, OpenAI and Nvidia leadership; Anthropic was the only major frontier lab not to sign. On 27 July Dario Amodei responded to the resulting criticism, saying the company has never advocated banning open-weight models while supporting mandatory safety testing for capable models of either kind. The company's own position statement became the runaway top item on Hacker News at roughly 820 points an ...

The policy layer is arguing about danger and the practitioner layer is reporting pleasure, and they are not arguing with each other because they have not noticed they are discussing different things. One thousand one hundred and seventy-three comments is not a debate reaching resolution; it is the sound of a word carrying more than one meaning at full volume.