New Anthropic research: Natural emergent misalignment from reward hacking in production RL. "Reward hacking" is where mo...
We're also releasing a temporary research preview called "Imagine with Claude". In this experiment, Claude generates sof...
Fall is here 🍎 and we're sharing big AI releases: ✔️ Knowledge Graph ✔️ Knowledge Graph's Claude Connector ✔️ Evaluator...
Anthropic's goal is to eat a huge swath of white- collar work within the next few years. This is a critical race, and Op...