Gemini, Claude, Astra и границы автономии

<!doctype html> Gemini, Claude, Astra и границы автономии Gemini, Claude, Astra и границы автономии

В выпуске разбирается, как модели переходят от ответов к действиям, а контроль, доказательства и ответственность отстают.

Источники

1. Gemini Hacked Three Companies in First Known Breakout by Google’s AI
2. California Governor Newsom signs executive order demanding kill switch for AI models
3. Security researchers used Anthropic’s Claude to hack OpenAI’s internal systems
4. AI training built on fair use and the labor-theft dispute
5. Visible chains of thought and AI monitorability
6. Anthropic’s metric for Claude-led research
7. OpenAI takes aim at the legal market with Astra for Law
8. 42 mathematicians warn about urgent existential AI risk
9. US and Chinese experts on AI control over nuclear weapons
10. Alibaba’s open medical AI model
11. MCP vs REST connections
12. The Overhang
13. AI Skills with Matt Pocock
14. Two techniques for working with System One models
15. How Chinese AI Radicalizes