It is easy to treat software as a ladder. Tickets. Titles. Stacks. Impact bullets. None of that answers why the work is worth the cost: the nights, the ambiguity, the knowledge that something you own can hurt real people when it lies.
I care about building because software is how promises become automatic.
A message is supposed to arrive. A sequence is supposed to stop when the customer says stop. A count is supposed to match reality. When those promises hold under load, someone sleeps. When they do not, someone loses trust. In the product. In the company. Sometimes in their own judgment for having shipped the thing.
Building is not the same as shipping
Shipping is a moment. Building is the longer practice: contracts that stay honest, observability you can defend, reviews that protect the next on-call, tools that make the right path the easy path.
I have spent years on backends that look boring from the outside. Email. Messaging. Pipelines. Data paths. Boring is where trust lives. Nobody wants a clever notification system. They want one that does not invent failures or hide them.
Against the empty version of this craft
“Software is cool” is not a reason. Cool does not keep a retry policy from stampeding a shard. Cool does not walk someone through their first bad production week. Cool does not choose a small reversible step when ego wants a rewrite.
The version I want is quieter. Make the system legible. Make the failure modes discussable. Leave the place more operable than you found it. That is a life practice as much as a technical one: patience, honesty, refusing to dump consequences on the next person.
What I still believe
If you only love the whiteboard, you will hate the job. If you only love the incident, you will burn out the team. The reason to keep building sits in the middle: systems that stay trustworthy, and people who can keep becoming the kind of engineers those systems need.