Automate Everything You Can

What to automate?

  1. Human instinct of boredom is the best indicator of where to automate. If something is boring, it is what to automate.
    • It's nature's way of pointing toward automation opportunities.
    • When you feel bored, your mind is telling you: "This task is beneath your cognitive capabilities."
    • The Boredom Principle
  2. Track your life and find the bottlenecks, places where takes most of the time. Prioritize them for automation, as Amdahl's Law

What can be automated better?

When to automate?

How long can you work on making a routine task more efficient before you're spending more time than you save? (across five years)

  • how often you do the task X how much time you shave off
50/day 5/day Daily Weekly Monthly Yearly
1 sec 1 d 2 h 30 m 4 m 1 m 30 s
5 sec 5 d 12 h 2 h 21 min 5 m 2 s
30 sec 4 w 3 d 12 h 2 h 30 m 2 m
1 min 8 w 6 d 1 d 4 h 1 h 5 m
5 min 9 m 4 w 6 d 21 h 5 h 25 m
30 min + 6 m 5 w 5 d 1 d 2 h
1 hour + 10 m 2 m 10 d 2 d 5 h
6 hours + + + 2 m 2 w 1 d
1 day + + + + 8 w 5 d

When not to automate?

  • One-off tasks
  • Highly creative work
  • Tasks requiring human judgment
  • When automation cost exceeds long-term benefits

The compounding effects

Find more examples

Leverage Dichotomy of Human vs Machine Thinking

As a last thing, the goal of automation is not efficiency for efficiency's sake, but to make more time for human "things": creativity, innovation, and human connection.

How to decide what to automate?

https://xkcd.com/1205/

https://xkcd.com/1319/

https://xkcd.com/1988/

Local graph
The Art of Unix Progr…Sync, Build & Deploy …The Principle of Leas…Automate Everything Y…