White hat, black box: AI’s next chapter

Anthropic’s unreleased "Mythos" AI model marks a shift toward cautious deployment as its advanced hacking capabilities pose significant cybersecurity risks. By granting exclusive preview access to systemically important entities like Apple and Microsoft, Anthropic aims to patch vulnerabilities—such as a 27-year-old bug in OpenBSD—before potential malicious use. This staged release model addresses compute constraints and intellectual property theft while potentially establishing a new regulatory standard for high-risk AI. Simultaneously, India’s political landscape is transforming as female voter turnout now surpasses male participation. Politicians are increasingly courting this demographic through direct cash transfer schemes, which currently consume significant state revenues, such as 10% of West Bengal’s receipts. While these programs provide immediate financial independence, they risk crowding out long-term structural investments in education and healthcare. Additionally, Senegal’s national football team faces high expectations for the upcoming World Cup despite recent continental controversies and the country's precarious national debt.