This is the most disjointed of his three books, but the first section alone makes it worthwhile. Tim breaks down his methodology to learning anything which has been incredibly useful to me for writing and marketing, and a number of other skills I’ve picked up bits and pieces of.
The next 4 of 5 sections are a fun cookbook, which help explain the concepts through practical exercises. They’re good if you’re new to this methodology, but if not, you don’t need them as much.
Averages are bad indicators because in extremistan / non-linear situations (see Antifragile) you get major outliers that destroy the average (Bill Gates walks into a bar, the average net worth goes up by a billion dollars).
What you study is frequently more important than how you study. You can spend a lot of time studying things you don’t need (learning “niece” in Spanish when you have no nieces).
Two lenses for viewing learning methods:
Use “DiSSS” and “CaFE” as frameworks for learning:
You can break deconstruction down into 4 pieces:
Reducing: Figure out the composite parts that you can break the skill down to. E.g. the Kanji characters of Japanese.
Interviewing: Find someone near the top, or who was at the top, that you can talk to about the skill.
Some example questions (for ultra endurance running):
Reversal: What would happen if you did things in the reverse of common practices? The best fires are built upside down, largest logs on the bottom building up to kindling. Try googling “backward” or “reverse” plus whatever it is that you’re trying to do.
Translating: Finding a common way to explain all the composite parts of the skill (e.g. “deconstruction dozen”) is a great way to quickly immerse yourself in the composite parts when talking to someone who is more skilled. Look for the “helping verbs,” the little tricks of any skill that let you pick it up very quickly.
Simple works, complex fails. Focus on the most important pieces that will give you as much fluency in the skill as possible. 100 well selected words give you 50% of a language.
Whatever the skill is, find the pieces that will give you the most leverage fastest.
It’s the burden on working memory that makes something easy or hard, and good sequencing can reduce the burden on working memory by stringing things together (see Power of Habit, Willpower Instinct).
Many effective chess players start by learning the end-game first.
There are many skills that are implicit, things the pros do without knowing it that they won’t think to teach you.
How do you make failure as painful as possible? Sticks work better than carrots because of loss aversion.
See Go Fucking Do It for an easy method.
The easiest way to avoid being overwhelmed is to create positive constraints: restrict whatever it is that you’re trying to do. Take advantage of Parkinson’s Lawinstead of having it work against you.
Create a prescriptive and practice one pagers:
Compress as much as possible to keep it all in your head.
Skill learning will tend to follow a curve like this (as does company growth, relationships, other things):
Based on the curve, you can forecast setbacks, plateaus, etc.
Spanish in one month forecast:
Take breaks regularly: we remember things better at the beginnings and ends of sessions, so more breaks = more benefits from the serial position effect.
Encoding allows you to take things you’re not familiar with and make them familiar, turn them into pieces you can attach to your existing knowledge.
Chunk things into more memorable pieces, instead of trying to remember every individual piece, or the whole thing at once.
Use a “memory palace” to remember things like numbers, cards, etc. (see Moonwalking with Einstein).
At this point, the book turns into a cookbook and the “meta learning” info is done. I’d encourage you to get it for yourself.