Authorities from seven Latin American countries, including Brazil, report difficulty accessing Anthropic's Mythos model and OpenAI's Daybreak project. In both cases, only authorities from the United States and Europe have access.
The models are used to locate vulnerabilities in systems and reinforce digital defenses, with speed gains of up to ten times compared to conventional work, according to engineers who had access to versions without limiters.
When contacted since Monday, OpenAI and Anthropic did not respond about possible partnerships with Brazil and other countries in the region. OpenAI stated that technical cooperation should be led by the American government.
