Design deliberate practice with AI
Choose one weak point, practise it at a manageable level of difficulty and decide what feedback will tell you whether the next attempt was better.
Design deliberate practice with AI when adding more time to practice is giving you the same result. The deliberate-practice skill narrows the task to one weakness and asks how you will recognise an improvement. A musician who loses the pulse during one transition can work on the transition without repeating the entire piece.
Feedback needs to arrive while you can still remember what you did. You might listen to a recording, count missed entries or ask a teacher about one bar. The next attempt changes one thing in response. Where there is no clear measure, the guide treats the available signal as an approximation.
What it does
- Isolate a recurring difficulty inside a much larger task.
- Choose a variation just beyond your current comfort level.
- Arrange feedback you can use immediately after an attempt.
- Compare a later attempt against the same visible criterion.
- Revisit a basic component when harder exercises stop helping.
Install and use
- Install the skill in your agent and describe a recent attempt and the point where it broke down.
- Ask, “I lose the beat during this transition. Help me design a brief exercise with feedback I can hear.”
- Try it, report what happened and choose a single adjustment for the next repetition.
FAQ
Source
Made by Amey Thakur and Archit Konde. Original skill: deliberate-practice, published on GitHub under the MIT licence.
View on GitHub