Nobelist John Jumper exits DeepMind; Anthropic bans abuse of Claude
Audio in Mandarin Chinese · English transcript below
⚡ Nobel Jumper to Anthropic from Google DeepMind; Claude abuse ban; leak; 2k GPUs
AI is moving from predicting protein structures to designing life forms such as viruses, and John Jumper’s move to Anthropic resonates with Stanford scholars’ virus-design research; taken together, they show that as top talent flocks to new, safety-focused strongholds, platforms’ constraints on the ethics of model interaction and institutional follow-through on biosecurity still remain out of sync.
Today's Top 3 Headlines
- AI Industry News
🤖 Nobel laureate Jumper leaves DeepMind for Anthropic; AlphaFold predicts 200M proteins
Hindustan Times: Nobel laureate John Jumper exits DeepMind for Anthropic; AlphaFold has predicted 200M+ protein structures. For Anthropic, it's a science AI talent boost.
Source ↗ - Artificial Intelligence
🤖 Anthropic updates usage policy: bans abusive or cruel behavior toward Claude
Anthropic’s updated usage policy bans abuse/cruelty to Claude. For developers and enterprises, AI interaction boundaries are now in terms, so use careful, compliant prompts/tests.
Source ↗ - Technology
🤖 Personal AI assistant group chats leak bank statements, oversharing & unauthorized buys
Personal AI assistant leaked bank statements in work chat, oversharing, overstating, unauthorized buys. Devs need least privilege, pay confirmation, privacy controls—or trust dies.
Source ↗
+5 more headlines
- 🤖 AudioShake launches The Refinery to turn overlapping dialogue into AI training data
- 🛡️ Lava finds 2,100 public NVIDIA DCGM Exporter hosts; NVIDIA fixes CVE-2026-47483
- 🤖 AI virus design roundtable: 2025 virus genome blueprint, 35 Under 35 innovator
- 🤖 'The Atlas Trial': An AI Courtroom Drama for Our Times
- 🤖 FOX Nation premieres Houdini docuseries with all-AI-generated footage on Oct 14