Algorithms to Live By

Algorithms to Live By is a very enjoyable and applicable book by Brian Christian and Tom Griffiths that explores how we can use knowledge from computer science to guide decisions in our lives.

Whilst surfing around for a synopsis I found this excellent summary, please take a look and come back afterwards for my own observations below.


There are many algorithms for planning the day. One thing that stood out was thinking about small and large tasks and how sometimes a small task can block the large task.

I think I like to do planning that involves planning the large things first and then doing the small tasks after the large ones. I think I also will talk with the rest of the office about doing this type of planning (which we already do to some extend).

For my personal life I also like this with regards to what to do when I get home. First do X, Y, Z (e.g. write, stretch, read) and only then think of what else to do.


You can see caching as the amount of information you have to keep in your head. You’re not very good at it. So write down almost anything. I always have my notebook with me or can make a note in Gmail so I think I’m quite good here, but it can be even better (e.g. by then sorting those notes to the relevant places). And also using Todoist more.

Optimal Stopping

The next few lines are a bit tongue-in-cheek. With optimal stopping you look at when to stop doing X to find the perfect Y. Dating can be an example here and in the book they mention that stopping at about 27 should be the right age (if I remember correctly). Guess around which age I met my girlfriend.


With regards to this algorithm I can say that I like how this makes you look positively at doing something again (exploit) because you liked it (e.g. eat at a specific restaurant) and not doing something new (explore).

For me the practical examples are friends (see many of the same ones again and again), foods (where I could explore more, also because I think there is value in the exploring itself), music (Spotify does a good job of balancing both).