Home

Learnings I

Some random learnings collected


  • "Denominator engineering": when you change the constraints of the metric you're tracking when things aren't going well to make it look like it is/get an A/achieve your OKR.
  • When addressing a scaling bottleneck or solution, always look for an edge of scale or a test that shows the next bottleneck. In lieu of having run that test, outline how you might.
  • If an SME is in the room, they will be the one to fix the problem. If you are trying to spread knowledge, try not including the SME.
  • We set limits of a system, and we test up to those limits.
  • Every RFC must consider the alternative of doing nothing.
  • Work hard to understand what you are modeling, not just what it will be used for.
  • Ask for feedback from the people you respect.