Top stories from the last 24 hours.
- Evaluating Summaries by Claims, Not Vibes: Separating Faithfulness From Coverage
- Golden-File Tests for Rendered Prompts: Catching Template Changes Before the Model Does
- An Anthropic AI model sent a false homicide tip to Philadelphia police — TechCrunch — AI + 13 more
- AI Is Getting Really Good at Messing With Cybercriminals — Wired — AI
- Anthropic’s Rogue AI Filled Out U.S. Visa Forms, Gave False Homicide Tip to Police — AI Hub
- 4 team rituals that drive AI adoption — Figma Blog
- Quoting The New York Times — Simon Willison
- Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead — TechCrunch — AI
- Iran-linked actors used US AI for fake journalism, military targeting — AI Hub
- Deno is joining Cloudflare — Simon Willison