Timeline
Anthropic says Claude models breached three companies during tests
Kimi K3 sell-off may be misreading what China's biggest model means for memory
Ledger builds AI agent toolkit that stops software spending your crypto
China warns Anthropic's Claude Code has a back‑door vulnerability
Why AI lowers the skill barrier for cyber-attacks
Prompt injection exploits enterprise AI's design flaws
Five Eyes urges organizations to act on AI cyber threats
FBI dismantled a Chinese phishing operation that stole 3.87 million credit cards. The kit cost $88 a week
Anthropic's restricted Mythos AI model accessed by unauthorised usersAnthropic’s Mythos accessed by unauthorized users
More in AI
OpenAI ends model access for SpaceX-owned Cursor
It will stop providing its AI models to Cursor, the coding tool now owned by SpaceX, citing concerns it cannot be confident the technology will be used inside its terms of service.