The case for boring tech
The most innovative feature in your product won't matter if your novel database falls over on Tuesday.
- PostgreSQL
- Simplicity
- Architecture
- Next.js
Every new technology brings an unwritten ledger of unknown failure modes. We choose boring, battle-tested foundations so our innovation budget is spent entirely on the user's problem.
in short
Key takeaways
- 01
Spend your novelty tokens wisely
Most startups only have three novelty tokens to spend before complexity collapses the velocity.
- 02
PostgreSQL solves 90% of data problems
Relational queries, JSON documents, full-text search, and pub/sub queues all live comfortably inside Postgres.
- 03
Debuggability is velocity
When something breaks at 2 AM, standard stack traces and readable logs beat esoteric distributed systems every time.
the details
The deep dive
The innovation budget
Dan McKinley's seminal essay on boring technology coined the concept of "innovation tokens". You get three of them. If you spend one on an obscure web framework, one on an experimental database, and one on a distributed graph store, you have zero tokens left to invent what actually makes your business unique.
We build on Next.js, TypeScript, and PostgreSQL. These tools have millions of person-hours of hardening, exhaustive documentation, and known operational characteristics.
Boring doesn't mean obsolete. Boring means you know exactly how it behaves when the disk fills up. ā Soren Patel, Staff architect
Postgres as your Swiss Army knife
Need full-text search? Before installing Elastic, use Postgres tsvector. Need key-value storage? Use JSONB with a GIN index. Need a lightweight job queue? Use SKIP LOCKED queries.
Consolidating on one primary data store drastically reduces operational overhead, backup complexity, and connection pooling friction. When your scale outgrows it, you can extract specialized services with real telemetry to guide the migration.