Author

Brian Christian

16 key ideas across 2 books, each explained in our own words.

Algorithms to Live By

IDEA 1
Computer science can guide daily choices
Christian and Griffiths argue that many everyday problems, such as finding a flat or organising a cupboard, resemble problems computer scientists have studied. Algorithms developed for computers can offer surprising advice for human life.
IDEA 2
Know when to stop looking
The authors explain optimal stopping, such as deciding when to stop searching for a flat. One rule suggests looking at about the first thirty-seven percent of options, then choosing the next one better than all before.
IDEA 3
Balance exploring and exploiting
The authors describe the explore-exploit trade-off: trying new options versus sticking with known favourites. Young people may explore more, while older people may exploit. The right balance, they say, shifts with how much time you have left to enjoy what you find.

IDEA 1
Machines may misread our intentions
Christian explains the alignment problem: making sure machine learning systems do what humans actually intend. As these systems take on more important roles, gaps between our goals and their behaviour can cause real harm.
IDEA 2
Data can carry bias
Christian describes how AI trained on biased data can make unfair decisions, such as in hiring or criminal justice. The bias comes from the examples it learns from. Fixing it requires looking at the data itself.
IDEA 3
Fairness is hard to define
Christian explains that there are several definitions of fairness, and they can conflict. Choosing between them involves values, not just mathematics. A system cannot satisfy every definition at once, so someone has to decide what counts as fair.

Go deeper with Brian ChristianAsk questions, argue with the ideas and practise using them, free.