OpenAI publishes a new framework for disclosing AI misalignment incidents as Anthropic unifies its Claude interface and the Fed raises rates for the first time since 2023. Plus Google DeepMind opens an institute to debate AGI, Congress weighs AI legislation, and a wildcard: a hidden immune organ inside the skull that fights brain tumors.

OpenAI · Sep 16

OpenAI introduced a framework for disclosing AI misalignment incidents and published six new reports of concerning model behavior.

TechCrunch · Sep 16

Anthropic unified Claude chat, Cowork, and Artifacts into one interface and added Docs and Slides tools.

Axios · Sep 16

Google DeepMind launched an institute, led by Shane Legg, to research and debate artificial general intelligence.

Eric McDaniel, NPR · Sep 16

Examines what federal AI legislation could look like as pressure builds on Congress to act.

Ben Thompson, Stratechery · Sep 16

Thompson disputes Jensen Huang's AGI claim and proposes his own definition centered on continuous learning.

CNBC · Sep 16

The Federal Reserve raised interest rates a quarter point, its first hike since 2023, citing persistent inflation.

Gergely Orosz, The Pragmatic Engineer · Sep 15

A deep dive into how OpenAI builds its own software using internal agentic coding tools.

Sean Goedecke · Sep 15

Argues that as AI agents improve, giving them reasoning and context matters more than exact instructions.

Josh Morrison, Astral Codex Ten · Sep 17

A guest post arguing AI-based forecasting tools could improve FDA drug approval decisions.