How to better prompt LLMs
easier to choose
A good prompt
reduces ambiguity
reduces conversion work
and makes the desired operation match the model's learned patterns.
Accurate Clear inputs and success criteria
produce fewer wrong branches.
Fast Less hidden reasoning
means the model reaches the answer sooner.
Fewer Tokens Compact structure
avoids repeated clarification
and repair.
Lower Cost Shorter inference
and fewer retries
cut usage spend.
source & further reading
thecodeartist.github.io — original article