There's a tax on novelty in infrastructure choices. The new framework, the new database, the new runtime — each promises better performance, cleaner ergonomics, or a more elegant mental model. Sometimes those promises are real. The cost, almost always underestimated, is the engineering hours you'll spend on the inevitable rough edges that mature technology has already smoothed over: the missing libraries, the documentation gaps, the community that hasn't yet hit the same bugs you will.
We have a rule of thumb at TRG: pick boring infrastructure on purpose. PostgreSQL instead of the new graph database for the project that has tables and columns. Next.js or Rails for the web app instead of whatever framework launched on Hacker News last month. AWS or Cloudflare for hosting instead of the experimental edge platform. Boring tech has the property that the answer to most questions is already on Stack Overflow, written down five years ago, by somebody who hit your problem first.
The interesting tech belongs to the part of your stack where the differentiation matters — your product, your customer experience, the work that's actually competitive. The infrastructure underneath should be unsurprising. A Postgres database that's run on Linux servers for thirty years is faster to debug at 2 AM than a database whose entire support community fits in a Discord channel. When something breaks, and it will, the question becomes: how quickly can we find someone who's seen this exact failure before? Boring tech wins that race every time.
The mistake we see most often isn't choosing the wrong boring technology. It's choosing exciting technology for the boring parts. The dating app that runs on a graph database to store user profiles. The blog that's been rewritten three times in the last 18 months to chase the latest framework. The internal tool that hasn't shipped a new feature in two quarters because the team is migrating to a new ORM. None of these is a technology problem. They're all priority problems, manifested as infrastructure choices nobody quite has the political capital to question.