There is a line in the first Iron Man movie that I think about almost every time I start a side project.
Obadiah Stane is yelling at one of his engineers. The engineer says what they want is impossible. Stane snaps back: "Tony Stark was able to build this in a cave! With a box of scraps!"
It is supposed to show how angry Stane is. For me it became a small rule for building things.
The Mark I was ugly
Look at the first suit again. It is not the shiny red and gold armor from the posters. It is grey, heavy and bolted together from pieces of the same missiles Tony used to sell.
It has flamethrowers on the arms. It walks slowly. It flies for a few seconds, then crashes into the sand.
And it works. It gets him out of the cave. That was the only job it had.
This is the clearest picture of a minimum viable product I know. Not a small version of the final thing. A rough thing that solves one urgent problem with what is already in the room.
You build with what you have
Tony did not have a lab in the cave. No assistant, no supply chain, no time. He had scrap metal, a few tools, and one person helping him.
When I start a new project, my instinct is to wait for better conditions. The right framework. A clean design system. A free weekend. A clearer idea.
Then I remember the cave.
Today the box of scraps is huge, honestly. Open source components, ready auth, hosting that deploys on push, AI agents that write the boring parts. A solo developer has more parts lying around than a whole team had ten years ago.
The problem is rarely the parts. The problem is deciding to bolt them together before they feel perfect.
Escape first, polish later
The Mark I had one goal: escape. Not fly elegantly. Not look good in a press photo. Just get out alive.
Every first version should have a goal like that. One sentence. Can someone book a court? Can someone pay? Can someone get the answer they came for?
If the first version does that one thing, it is a success, even if it crashes into the sand right after.
I have killed more projects by polishing them than by shipping them too early. Nobody ever used the settings page I spent a week on. People did use the ugly form that just worked.
The garage comes after the cave
What I really like is what happens next. Tony gets home, and he does not stay with the Mark I. He goes to his garage and starts again.
The Mark II is still not the final suit. He tests it in small steps. A little thrust on one hand. Then both. Then a short hover that throws him into the ceiling. His robot helper sprays him with a fire extinguisher whether he needs it or not.
Then he takes it up high and finds a real problem. The suit ices up at altitude and drops out of the sky. So the next version gets a new alloy that fixes it.
That is the loop. Build. Test. Find the thing that breaks in the real world. Fix that one thing. Repeat.
You only find the icing problem by flying. No amount of thinking in the garage would have shown it.
Scraps are a feature
There is a quiet advantage to building from scraps. You cannot overbuild. Your limits make the decisions for you.
When you have everything, every feature looks possible, so you try to do all of them. When you have a box of junk and a deadline, you build only what matters.
I try to fake this sometimes. One weekend. One page. One problem. No new tools allowed. It feels like a constraint, but it is really a gift.
So, build the ugly suit
I am not saying be careless. Tony was a genius, and the cave suit was still smart engineering. It was just honest about what it had to do.
So that is my reminder for the next idea. Build the ugly version first. Make it escape the cave. Then go to the garage and make it fly.
The shiny red and gold one can wait. It always comes later anyway.
