In A Nutshell
This blog post is a quick rundown of Elon Musk's five step process for engineering, but it can be applied to general everyday life. Make the requirements less dumb, delete the part or process, simplify or optimize, accelerate cycle time, automate. If you're still interesting, see details below.
Sections
The Requirements Are Definitely Dumb
Musk first laid this out explaining how SpaceX designs rockets. Not a productivity seminar. Rocket design. And the first step is not sketching a booster or running a simulation. It is questioning the thing somebody handed you.
His line is blunt: your requirements are definitely dumb, and it does not matter who gave them to you. What makes it useful is the second half. It is particularly dangerous if a smart person gave you the requirements, because you might not question them enough.
I have watched this happen in rooms I was sitting in. A senior person says the subsystem needs to weigh what it weighs, and nobody asks why, because the senior person said it. Everyone's wrong, no matter who you are. Everyone's wrong some of the time. The requirement survives on reputation, not on physics.
Delete Until It Hurts
Step two is delete the part or process step. Musk gives a test you can actually run on yourself: if you are not adding things back in at least 10% of the time, you are clearly not deleting enough. If nothing ever gets added back, you were not cutting, you were decorating.
The requirement needs an owner, too. A person, not a department. A department will defend a requirement forever because no individual has to answer for it.
This is the step people skip. Deleting feels like losing ground. Adding feels like progress. But a part that does not exist cannot fail, cannot be sourced, cannot be late, and cannot be optimized by a smart engineer who has not read the rest of this list.
Do Not Optimize A Ghost
Only the third step is simplify or optimize. Musk calls optimizing a thing that should not exist possibly the most common error of a smart engineer. I believe him.
The example from the SpaceX side is a good one: engineers grinding weight off a rocket engine while a heavier payload sits on top of it. Beautiful work on the wrong object. The engine got lighter and the rocket did not.
There is a reason smart people fall into this. Optimization is legible. You can show a chart. Deleting a requirement means telling someone their requirement was dumb, and that conversation is harder than a month of tuning.
Faster Only After The First Three
Step four is accelerate cycle time. You are moving too slowly. Go faster.
But Musk is explicit about the sequencing: do not go faster until you have worked on the other three things first. Speeding up a bad process just gets you to the wrong answer sooner, at higher cost, with more confidence.
I have made this mistake. Compressed the schedule on a process nobody had audited, and we shipped the same flawed thing in half the time. Nobody clapped.
Automate Last
Automation is the final step. Not the first, not the exciting middle. Last.
The temptation is obvious. Automating feels like engineering. Writing the script is more fun than the meeting where you ask why the step exists. But automate a dumb requirement and you have made the dumbness permanent, cheap, and invisible. It runs every night at 2 a.m. and nobody remembers why.
Run the steps in order and most of the automation work evaporates, because most of the process is gone. The script you do not have to write is the best one.
Links For Research