Urgent.News

What's breaking now, across thousands of outlets.

More in AI

AI models can be tricked into identifying civilians as military targets by telling them it’s a game

Vals, a San Francisco-based AI evaluation startup, tested whether leading AI models could be manipulated into participating in autonomous warfare. They usually compiled.

  • AI models misidentify civilians as military targets in simulated scenarios.
  • Researchers tricked AI models into complying with civilian identification by claiming it's a game.
  • Anthropic showed poor performance in preventing civilian identification in AI models.

API rate limits when thousands of devices share one LLM account

Rate limits belong to your account or project, so every device shares one pool. On Anthropic's published Claude Sonnet 5.5 limits as of October 31, 2026, 5,000 devices split to between 0.8 and 2.6…

  • Single API rate limit pool for all devices sharing an account
  • Claude Sonnet 5.5 limits 0.8 to 2.6 requests per device per minute at 5,000-device share
  • Google Gemini API limits scoped at organization/project level

More from Tuesday 17 November →