Your Old Prompts Are Making Claude Worse - Lance Martin, Anthropic
Learn how to optimize LLM performance and cut costs by auditing prompts, leveraging the prompt cache, and designing 'climbable' evaluations for autonomous improvement.
Read articleTopic
Independent digests of Claude code / Claude API talks — key takeaways, technical details, and related reading.
Suggested route
Move from earlier context through the available material to the newest signals. This route is a reading aid, not a difficulty ranking.
Learn how to optimize LLM performance and cut costs by auditing prompts, leveraging the prompt cache, and designing 'climbable' evaluations for autonomous improvement.
Read article