Ultralearning

Главные выводы

• Hard skills can be learned fast when you design the project yourself and go all in
• Spend the first 10% of your time mapping what to learn and how experts learn it
• Practise the real thing in the real setting, because skills transfer poorly from classrooms
• Isolate your weakest component, drill it, then put it back into the whole
• Seek harsh feedback early and keep experimenting once you hit a plateau

Внутри этого бита

1. Learning hard things on your own terms

   Scott Young once set himself a strange challenge: pass the exams for MIT's four-year computer science degree in 12 months, using free online lectures and no enrolment. He did it.

2. Draw the map before you walk

   Most learning projects fail in the first week because nobody planned them. Young's first principle is metalearning: research how the skill is actually learned before you start.

3. Do the real thing, then drill

   The principle Young calls directness says to practise the skill in the situation where you will use it. Skills transfer far less than we assume.

4. Test yourself and demand feedback

   Young builds on the same memory science that underlies flashcards. Retrieval, pulling knowledge from your head instead of reviewing it, is the foundation of remembering.

5. Build intuition, keep experimenting

   Why could physicist Richard Feynman see solutions others missed? Young's answer is that deep intuition is not a gift but the product of a specific habit: refusing to move on until you truly understand.

6. Put it to work

   Pick one skill you have been meaning to learn and give it a clear finish line, such as holding a 10-minute conversation or shipping a small app. Spend two hours this week on metalearning: find two people who have learned it and ask how, then list the concepts, facts, and procedures involved.