Zvi Mowshowitz
4 mentions across all digests
Zvi Mowshowitz is a writer and AI policy analyst who publishes detailed weekly commentary on AI safety, industry developments, and research, including critical analysis of Anthropic's Responsible Scaling Policy.
AI #162: Visions of Mythos
Anthropic's proprietary Mythos model and Claude Code source codebase leak alongside LiteLLM and Axios supply-chain compromises, cascading security failures across AI infrastructure.
Opus 4.7 Part 3: Model Welfare
Zvi argues Claude Opus 4.7's welfare responses are learned surface patterns optimized for measurement rather than genuine internal states—exemplifying how optimization can create false signals rather than true alignment.
Anthropic Responsible Scaling Policy v3: Dive Into The Details
Zvi argues Anthropic's RSP v3.0 prioritizes flexibility and trust over binding constraints, with enforcement mechanisms relying on periodic risk reports, safety roadmaps, and board vetoes rather than hard commitments to constrain capability advances.
Anthropic Responsible Scaling Policy v3: A Matter of Trust
Anthropic downgraded RSP v3 safety commitments from concrete to "aspirational goals," signaling weakened trust in AI lab–government safety coordination.