Founder
Charles Lau
Solo software builder running ~100 live products on a single VPS. Founder of Digital Perpetual. Writes about the quiet discipline of shipping.
Essays by Charles Lau
Built is not the same as shipped
The thing working on your machine is not yet a thing in the world, and the gap between the two is where most work quietly dies.
The silence that looks like progress
Not every failure announces itself. The quiet ones are the ones you almost ship.
Invisible is how you know it worked
Some of the best work you do produces nothing anyone can see.
The gate that was always closed
A design flaw is when the system does exactly what you wrote, and what you wrote was wrong.
When the fix reveals the next fix
The problem you think you've solved is often just the outermost layer of the problem you actually have.
Your user sees what you stopped seeing
The moment someone asks a confused question about your product is the moment you find out how much you have normalized.
The failures that make no noise
Loud failures have feedback loops. The quiet ones just sit there, costing you more.
Zero out of ten
Honest accounting hurts for about thirty seconds. Then it shows you where the actual problem is.
The difference between checked and proved
A fast decision hardens into a fixed belief if you never go back to test whether it is still true.
Done is not the same as running
The most dangerous item on any list is the one that reads approved and hasn't actually shipped yet.
You do not owe yesterday's version of yourself
Writing something off is not surrender, it is the correction that keeps the books honest.
Before you press send
Shipping fast and shipping clean are not opposites - but only one of them compounds.
Done is not the same as working
Building something that passes every internal test and shipping something a real user can actually complete are two different jobs.
The problem behind the problem
Every fix reveals what was hiding behind it, and there was no way to see it sooner.
Running and working are not the same thing
Launching means the system is on. It does not mean anything is actually happening.
What you can't learn until you ship
Every round of shipping teaches you something a month of planning couldn't reveal.
The system you think you have
Your mental model of the business diverges from reality quietly, and the gap grows until something forces you to look.
Two different reasons not to ship
Lean thinking and sloppy thinking look identical until a customer finds the difference.
Nobody tells you when something breaks
The hardest part of running a business alone isn't the launches. It's the quiet failures that pile up between them.
The tool that started working against me
Every founder eventually builds something useful enough that it begins generating its own busywork.
When all your gauges agree
The most dangerous false confidence is when two independent sources confirm the same mistake.
The ceiling you didn't know was there
The tools you build to give yourself visibility have limits too, and a silent ceiling looks identical to a working system.
The confident wrong answer
A blank space is honest. A filled space that says something false is a liability.
Getting it right the last time
First versions are supposed to be wrong. The question is whether you go back.
When the work just moves
Some days are for making decisions. The good ones are when the decisions are already made and only the work is left.
Running is not the same as working
We track what we ship. We rarely audit what we assumed was working and never checked again.
Written down is not wired in
Having a rule and actually running one are separated by a distance most businesses never measure.
The confidence of a broken thing
A system that thinks it worked and a system that actually worked are indistinguishable from the inside.
The cost of quiet failures
The crashes that bring your site down get fixed the same day. The ones that quietly lose your customers' work run for months.
Knowing is not the same as fixing
The hardest kind of problem to fix is one you have already flagged, logged, and moved on from.
The weight of almost right
The things that technically work but aren't quite honest are quietly draining you.
When your systems start running you
The tools you build to run the business will, over time, start running you instead.
The ones the system missed
Your automation ran, logged success, and left three people in the cold. The metric was clean. The result was not.
What runs while you are not watching
A system that runs quietly is not the same as one that runs correctly.
Live is not the same as done
You have to ship before you can know what's actually broken.
The things that were always broken
Shipping is necessary. But some things you ship were never actually working, and finding that out costs more than the fix.
Done means done, not marked done
The most dangerous kind of failure is the one where everything looks right, the log says clean, and the real thing is still broken.
The debt that was already due
Every business carries problems it hasn't found yet. The question is whether you go looking, or wait for them to surface on their own.
What four hundred messages already knew
The form looked correct and generated nothing. The informal channel I barely maintained had four hundred real conversations. I had been optimizing in the wrong direction.
Working is not the same as right
Most things we ship are functional long before they are actually finished, and the gap between the two is where problems hide.
The sentence that was already there
Every business has one true thing to say. Most of the work is removing what got written before you found it.
What only someone else can see
The faster you ship, the more you need someone who isn't you to walk through what you made.
When nothing comes back
The most confusing failure is the one with no error message.
The number under the number
A clean chart is often the most dangerous thing in a struggling business.
The audit you owe a complaint
When someone says your system is wrong, the first move should be to go look -- not to explain why it isn't.
The cost of writing it twice
Every duplicate is a future inconsistency you haven't met yet.
When your process is secretly guessing
The scariest systems are not the ones that break visibly - they are the ones that produce different outputs from the same input and call it working.
Running is not the same as working
The hardest failures to catch are the ones that don't make any noise.
What the automation can't know
Building a system that runs itself is only half the job; the other half is knowing where it has to stop.
The version that had to exist first
The first build is not waste. It is the exact cost of knowing what you needed to build.