Every technology choice you make today will be wrong eventually. Not because you chose badly, but because technology landscapes shift, business requirements evolve, and the context that made a choice correct in 2024 may not exist in 2029. The goal is not to choose technologies that never need to change. The goal is to choose in ways that preserve your ability to change.
Why Technology Choices Age
Technologies age in predictable ways. Some become obsolete as better alternatives emerge. Some lose maintainer support and accumulate security vulnerabilities. Some get acquired and change pricing or development direction. Some simply fall out of the talent market, making it increasingly expensive to hire engineers who know them. The right question when evaluating a technology is not just "is this good today?" but "what does the support and adoption trajectory look like over the next five years?"
The Trap of Novelty
The engineers who choose novel technologies are often rewarded in the short term — the technology is exciting, recruitment is easier because curious engineers want to work with new things, and the team feels like it's at the leading edge. The cost comes later, when the technology matures in an unexpected direction, when early architectural decisions it required turn out to be wrong, or when the talent pool for the technology is too thin to hire from at scale.
“Choosing proven technologies is not conservative. It is the risk management decision that preserves your ability to hire, scale, and change.”
Building for Replaceability
The principle that makes technology stacks resilient over time is the same principle that makes good software: separation of concerns. When your business logic does not depend directly on your infrastructure choices, you can replace infrastructure without rewriting logic. This requires discipline — it is always easier to let infrastructure choices leak into business logic — but it pays off every time the infrastructure needs to change.
The Evaluation Framework
- Maturity: how long has this technology been in production use at companies similar to ours? Is it past the phase of major breaking changes?
- Ecosystem health: is the community growing or contracting? Are major contributors investing in the project?
- Talent availability: how deep is the hiring pool? Is it getting deeper or shallower?
- Vendor risk: if this is a commercial technology, what happens to our stack if the vendor changes direction, raises prices, or is acquired?
- Migration path: if this technology needs to be replaced in five years, how hard is that migration? Is the interface well-abstracted?