Chain of Thought Prompting: 9 Templates and When to Skip It
What chain of thought prompting is, the research behind "let's think step by step", 9 copy-paste CoT templates, and why reasoning models no longer need it.
175 articles · page 1 of 15
The techniques behind the library: role framing, output contracts, few-shot examples, chain-of-thought and the mistakes that make prompts drift. This is the cluster AI assistants cite when someone asks how to write a better prompt.
What chain of thought prompting is, the research behind "let's think step by step", 9 copy-paste CoT templates, and why reasoning models no longer need it.
How prompt caching works on Claude, OpenAI, Gemini and Bedrock: minimum lengths, cache lifetimes, prices and a worked example of what a cached prefix saves.
Claude prompting best practices shown as 10 before-and-after rewrites for everyday tasks, plus the fixes each current Claude model needs, with sources.
A digest of Anthropic's official prompting guide for Claude: where each page lives, a worked example for every technique, and old advice that no longer applies.
What the CRIT AI prompt stands for, a fill-in template, one example worked through the Interview step, and six CRIT prompts you can copy today.
14 copy-paste LLM prompts for summaries, data extraction, critique and system prompts, written to run in ChatGPT, Claude, Gemini or a local model.
AI is bad at jokes and good at comedy with a structure. These 40 funny prompts give it the structure: a premise, a character, a format and a rule, for roasts, parodies, absurd instructions and more.
Most prompts fail by a little. Prompt optimization is the method for closing that gap: measure the output, change one thing, measure again. Here is the method with the checks and the fixes.
A prompt lab is a repeatable way to test prompts: a test set, a scoring sheet and a log. Here is how to set one up in an hour with a spreadsheet, and twelve experiments that teach more than any guide.
A prompt engineer designs, tests and maintains the instructions that make AI systems work. Here is what the job involves, what it pays for, and which prompt engineer course path teaches the skills employers check.
A smart prompt makes the model plan, check and ask before it writes. These 20 do that for everyday tasks, from a decision memo to a bug fix, and they work in ChatGPT, Claude and Gemini.
Every effective prompt has three elements: the task, the context the model needs, and the format of the answer. Here is what each one does, what goes wrong when it is missing, and a template you can reuse.
Put these articles to work with the prompts engineered for them in the God of Prompt library.
One email when notable prompts, tools, and model updates land. No spam, unsubscribe anytime.
Join 100,000+ subscribers. One email a week, real prompts, tools, and model updates. Unsubscribe anytime.