No Computer


I can’t remember the last time I spent more than a few days away from the computer. Last week I went on vacation and decided to take the opportunity to do an experiment in being offline.


There’s a concept called bracketing a problem. The idea is that in a world of uncertainty, we will get more knowledge by asking “what would be too little?” and “what would be too much?” than by simply asking “what is the right amount?” Once we have found the limits, we have bracketed the problem and we know where we stand 1.


Before starting this experiment, the question I had was: What will it feel like to not be connected and using screens? As the week went by, I realized that it didn’t make sense to treat work and leisure with the same hypothesis. My initial question forked into two semi-retrofitted hypotheses:

  1. I won’t hit the lower bound for leisure.
  2. I will hit the lower bound for work.


First, this was not a controlled experiment. The whole week I spent offline was away from home, and the main purpose of the trip was leisure.

The rules were quite simple:

  1. No computer.
  2. No smartphone except as necessary (minimal communication).

The experiment started Sunday/Monday at midnight and stopped the following Monday.


I didn’t bring a computer so it was impossible to use one. Phone I had turned off all the time except for two separate one hour periods when I periodically coordinated meeting up with someone. My iPhone had more than half its battery left at the end of the week, without charging it.

What follows are notes and observations from the experiment:

In summary, I didn’t hit the lower bound for leisure, and for my work (primarily programming) I did hit it.


I was surprised at how few things I missed. At some point I would probably begin to miss being in the loop, being able to go down Google and Wikipedia walks, and talking to people from all over the world, but I didn’t hit that point during a week.

I divided the initial question into one of work and leisure. Leisure I barely missed at all, as I just read books and did non-computer things. Work was different: without being able to look up things and code things on a screen - which, unlike mathematics, is the natural medium code is created in - I found myself unable to work on any programming. Initially my ambition had been to do some coding on a legal pad, but this proved to be difficult. It is possible this is something that can be exercised, but probably only for small, self-contained problems.

It was also easier than expected. I suspect this to a large degree depends on the state-dependent or context-dependent memory effect. It is much harder to remove old habits than to form new ones in a new environment. It is also much easier to just cut computers out completely than to use them only for dedicated tasks. Back in the real world, where we use computers every day, there’s a spillover effect - when we are working at a computer it’s easy to get into leisure-mode, at least temporarily. This is a struggle that a lot of people recognize, and it’s not a straightforward problem.

When I talked to fellow programmers about this experiment I was surprised by how positive they were about this. It’s as if they all secretly wish they could sneak away from the computer and just be away for a while.

Maybe not being connected all the time will develop into a form of luxury. Instead of asking when the last time someone travelled or took a vacation, the question we will ask in the future is: when was the last time you spent a week offline?

Thanks to Margo Kulkarni for reading drafts of this post.

(If you liked this, you might enjoy The Watsi Experiment.)

  1. One of the earliest references to this idea is Aristotle’s golden mean.↩︎