Reading8 min

Why Claude is Different

The model card and the safety-first design philosophy that distinguishes Claude from other frontier AIs.

The model card and the safety-first design philosophy that distinguishes Claude from other frontier AIs. https://www.anthropic.com/claude 8 min read
Reading15 min

How to Prompt Well

Anthropic's own guide to writing prompts that get the best out of the model.

Anthropic's own guide to writing prompts that get the best out of the model. https://docs.anthropic.com/en/docs/build-with-claude/prompt-engineering/overview 15 min read
Reflection15 min

Claude in Practice

Reflection Prompt

Pick one task you do every week — research, writing, planning. Run it through Claude. Where does it help, where does it not, where do you need to push back? Write down what you learned.

Reading12 min

Limits, Bias, and Hallucination

Where the model is unreliable and how to catch it. Includes Anthropic's public research on alignment and interpretability.

Where the model is unreliable and how to catch it. Includes Anthropic's public research on alignment and interpretability. https://www.anthropic.com/research 12 min read
Reflection10 min

Building a Practice

Reflection Prompt

Most people use Claude for one-off tasks. The compound returns come from using it as a steady thinking partner. Set up one regular use — daily writing review, weekly research synthesis, fortnightly strategic reflection. Stick with it for a month.