Justin Baird
Back to journal

Hardware Is Unforgiving, and That's the Point

Fixing an amplifier suspended 100 meters above a Broadway stage taught me what shipping hardware really means: get it right, because the physical world does not grade on a curve.

HardwareEngineeringCraft

My first job out of university was as a research and development engineer at Meyer Sound in Berkeley, California. This was many years ago now, but it taught me the single most important lesson I carry about building physical things: with hardware, you have to get it right.

At Meyer Sound I was building loudspeaker systems, and I mean the whole thing. Power amplifiers, power supply designs, wooden enclosures, the speakers themselves right down to the motor diaphragms, and then all the active circuitry and electronics on top. Real hardware boards. Real circuits. The kind of filters and component choices where one wrong resistor doesn't just break your prototype, it breaks the thousands of boards already shipped and sitting in venues all over the world.

I learned that reality in the most literal way possible. We had a problem with an amplifier that lived about 100 meters above the stage at The Lion King on Broadway. There was no clever remote fix. I flew to New York, went backstage in the theater, and we physically pulled the amplifier out of a speaker suspended 100 meters in the air, fixed it, and put it back. When your product is hanging above a live Broadway stage, you find out very quickly what "it shipped out the door" actually means.

That's the thing people underestimate about hardware. You can't just rely on the spec sheet. A component's datasheet tells you one thing, and then you put it in the actual circuit and it behaves differently. Getting something to work on the bench is only half the job, the other half is making it work in production, reliably, across thousands of units, in the messy physical world. Every resistor, every capacitor, every op amp has to be specified and chosen with the whole bill of materials and the whole manufacturing run in mind.

I want to be careful here, because I'm not saying hardware is more valuable than software. Making an app work is real work and real value. But they are two genuinely different worlds. When you type something on your computer and the app works, you can iterate endlessly. When you ship hardware, it's out there. If it's not right, you're going to know about it, and it can become a world of pain.

That pressure sounds like a burden, and sometimes it is. But it's also exactly what makes hardware so rewarding. The discipline it forces on you, the insistence on getting every detail right before it leaves your hands, is a kind of craftsmanship. Early in my career it drilled into me a standard I've never let go of: don't ship it until it's actually right, because the physical world doesn't grade on a curve.