"Caveman" prompting is a viral technique for Claude AI.
Write short. No filler words. No pleasantries.
Just direct instructions.
Goal: reduce tokens and cut costs.
Why it works:
LLMs read tokens, not sentences.
Long sentence → more tokens → higher cost
Short sentence → fewer tokens → lower cost
Example:
"I understand your request and will explain in detail"
vs
"Explain now"
Same meaning. Far fewer tokens.
Less tokens. More efficiency.
Longer conversations.
Repo: github.com/JuliusBrussee/…
#AI #LLM #PromptEngineering #Claude
English
