Essay
Why the best tools feel smaller
Beginners want simplicity because complexity is frightening. Experts want it because complexity charges rent.
-
01
Power
New layers make more futures possible.
-
02
Rent
Every layer must be learned and maintained.
-
03
Fit
Use only the power the present problem can justify.
The machine pays the creation cost. You pay the maintenance cost.
The second simplicity
Technology follows a funny curve. When you know very little, you want simple tools. When you know more, you want power and options. When you know a lot, you often want simple tools again. The second simplicity is different: the expert is counting costs other people do not see.
Experienced builders become enthusiastic about boring things—plain files, one process, a small command-line tool, a database that lives in one file—not from nostalgia, but because every layer adds cognitive, upgrade, debugging, hiring, and security rent.
Power you do not need
A small internal application can begin with a framework, build pipeline, API layer, authentication service, cloud database, containers, orchestration, monitoring, analytics, and several environments before the first customer pays. Each choice can be reasonable. Together they may form a system built for a hypothetical future instead of the present problem.
Simplicity is not religion. It is matching power to the problem. A highly interactive product may need a complex client architecture. A tool used by sixty colleagues probably does not need an architecture for fifty million simultaneous users.
Compression or postponement
Good simplicity is achieved, not primitive. A strong abstraction compresses understanding and lets you forget details safely. A weak abstraction postpones understanding until the day the details matter most.
One-button deployment is wonderful. If nobody knows what was deployed, where the data lives, how backups work, or what happens when the provider disappears, the button did not remove complexity. It moved complexity beyond the team's field of vision.
Attention is the scarce resource
Small teams have a special reason to prefer smaller tools. Every additional technology is a claim on shared attention. Your stack is partly an org chart written in software; if it assumes five platform teams and you have one engineer, you imported a company larger than your company.
AI can generate adapters, migrations, services, and state layers quickly enough to produce complexity faster than humans can understand it. Delay irreversible complexity. Choose the smallest architecture that provides a clean path to the next stage. For a startup, attention is usually more scarce than compute.