@shayneredford
AI research @ Anthropic. Prev: MIT Tech Review 35U35, Data Provenance Initiative, MIT, Google, Apple, Stanford. 🇨🇦 Views my own.
AI
The project's update says the catalog includes sources beyond those it used and asks readers to flag omissions.
AI
Anthropic says it costs around 75% less to run on average than Haiku 4.5 and improves coding and computer use.
AI
One critic argues that keeping disagreements internal can breed overconfidence about safety; another points to public data and researcher access as tools for independent scrutiny.
AI
MIT researcher Longpre appeared on a CBS Morning News Tech Watch segment.
AI
Anthropic says its model produced the first end-to-end computer-checked version of the theorem.
AI
Updated model targets coding, reasoning, and long-running agentic tasks.
AI
Public platform draws from consented conversations to measure actual AI assistant use.
AI
Recent MIT PhD graduate moves to San Francisco for role focusing on pretraining and AI societal impacts.