OpenAI and Anthropic agree to send models to US government for safety evaluations https://t.co/BjQV1ypmPI https://t.co/Qcvafg9BrH on https://twitter.com/AcerboLivio/status/1829282134726443250